CubeSat Communications: Recent Advances and Future Challenges

08/26/2019 ∙ by Nasir Saeed, et al. ∙ 0

The research in the emerging space industry is becoming more and more attractive, given the increasing number of space-related applications. One primary entity of current space research is the design of miniaturized satellites, known as CubeSats, due to their numerous applications and low design and deployment cost. The new paradigm of connected space through CubeSats enables a wide range of applications such as Earth remote sensing, space exploration, and rural connectivity. CubeSats further provide a complimentary connectivity solution to the pervasive Internet of things (IoT) networks, leading to a globally connected cyber-physical system. This paper presents a holistic overview of different aspects of CubeSat missions, and provides a thorough review on the topic, both from academic and industrial perspectives. We further present the recent advances in the area of CubeSats communications with an emphasis on constellation and coverage issues, channel modeling, modulation and coding, and networking. The paper finally identifies several future research directions on CubeSats communications, namely Internet of space things, low power long range networks, machine learning for resource allocation in CubeSats, etc.

READ FULL TEXT VIEW PDF
POST COMMENT

Comments

There are no comments yet.

Authors

page 6

page 13

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

I Introduction

The commercial race in the space industry is igniting, leading to an active new space economy. According to the report of Morgan Stanley, the expected revenue of space industry will rise to $22 billion by 2024, and to $41 billion by 2029 [1]. The space businesses are especially growing fast with the development of small satellites, because of their relatively low cost of deployment. Additionally, these small satellites are deployed in low earth orbit (LEO), thus providing low latency communications [2]. Due to the development of such small satellites, space technology is becoming cheaper, closer, and smaller, which is reviving space industry by providing various new applications such as space observation, earth observation, and telecommunications. These appealing applications of small satellites lead the top tier companies such as Google, SpaceX, OneWeb, and Facebook to investigate the use of such satellites at a low cost instead of using the conventional LEO satellites, to provide earth monitoring, disaster prevention, and connectivity to the internet of things (IoT) devices in remote areas.

Besides the economic space race in industry, academic researchers have rigourously joined the research of developing small satellites, which are often classified according to their weights, i.e., femto (less than

kg), pico ( kg), nano ( kg), micro ( kg), and mini ( kg) [2]. Among these satellites, pico-satellites, also known as CubeSats, emerge nowadays as the most popular ones. CubeSat program was initiated at Stanford university in 1999, to build a low cost and low weight satellite. Standards were defined thereafter to built these satellites in a cubic structure (thus the CubeSat terminology), with mass of 1.33 kg per unit, cost of less than $1000, low power consumption, and commercial of-the-shelf components. CubeSats have thereafter come to the center of attention for they can accomplish many scientific experiments for educational and institutional purposes due to their tiny structures [3]. In fact, there are more than thousand different CubeSat missions that have been launched over the past 20 years [4]. These missions can be categorized into four major fields, namely communications, earth remote sensing, space tethering, and biology [5] [6]. Few examples of these missions are included later in this survey so as to best illustrate their capabilities.

Recently, both academia and space industries have extensively investigated the applications of CubeSats as a means to provide global connectivity to the users around the globe. This has led to produce diverse projects and services in this field. For instance, KEPLER Communication launched their KIPP CubeSats in 2018 to provide connectivity to the users at the north and south poles. KIPP is the first Ku-band 3U CubeSat mission that offers 40 Mbps of data rate with 60 cm diameter of very-small-aperture terminal (VSAT) [7]. Analytical Space launched Radix mission to enable high-speed downlink communication using optical links. This mission started the transmission of data soon after it was deployed and lasted for six months. Radix mission consisted of 6U CubeSats, and its primary purpose was to test laser capabilities in downlink communications [8]. Another well-known mission for CubeSats communications is GOMX-3, supported by the European space agency (ESA). GOMX-3 consisted of 3U CubeSats, and was a successful mission which acquires signals from the worldwide aircrafts. [9].

Ref. Launch Year Mission Name Mission Type Size Frequency Band No of CubeSats Status
[7] 2018 KIPP Providing global connectivity 3U Ku-Band 2 Operational
[8] 2018 Radix Optical communication test 6U Optical 1 Successfully completed
[9] 2015 GOMX-3 Aircraft signal acquisition 3U X-Band 1 Successfully completed
[10] 2013-2018 Lemur-2 Weather forecasting 3U - 100 Operational
[11] 2011 DICE Ionosphere Monitoring 1.5U UHF-Band 2 Successfully completed
[12] 2003 QuakeSat Earthquakes forecasting 3U UHF-Band 1 Successfully completed
[13] - OLFAR Low radiations analysis - VHF 50-1000 Under review
[14] 2010 RAX Space weather forecasting 3U S-Band 2 Successfully completed
[15] 2018 MarCO Relaying for deep space 6U UHF and X-Band 2 Not Operational
[16] 2017 ISARA Bandwidth communication test 3U Ka-Band 1 Operational
[17] 2015 AeroCube OCSD Optical communication speed test 1.5U Optical 2 Successfully completed
[18] 2017 ASTERIA Attitude control test 6U S-Band 1 Operational
TABLE I: List of a few well-known CubeSat missions

Besides their communications missions, CubeSats hold a good number of missions dedicated for scientific understanding and prediction of the Earth environment. Weather prediction, climate change, and disasters monitoring are the most common applications of Earth science missions. For instance, Lemur-2 is a LEO constellation CubeSat mission by Spire [10]. Lemur-2 satellites consist of two payloads for weather prediction and tracking of ships. Dynamic Ionosphere CubeSat Experiment (DICE) is another well-known mission led by Utah State University to monitor the Earth’s ionosphere [11]. DICE CubeSats measure the in-situ plasma densities using two Langmuir probes for geospace storm-time features identification. QuakeSat (3U CubeSat) mission was launched in 2003 to forecast earthquakes. QuakeSat primarily carries magnetometer housed in a 60 cm telescope collecting the fluctuation measurements [12]. QuakeSat scans the global changes in extremely-low-frequency-electromagnetic (ELF) waves, which are believed to precede a seismic activity.

CubeSats applications are also popular at universe exploration and space science missions, which focus on expanding the scientific knowledge on astronomy, Heliophysics (space weather), and planetary science. For example, analyzing the cosmic background radiations below 30 MHz is exceptionally challenging, since it requires space-based large aperture radio telescope, which are often quite sensitive to the ultra-long waves in space. Hence a distributed system consisting of a swarm of 50 or more CubeSats called orbiting low-frequency antennas for radio astronomy (OLFAR) in the lunar orbit is used to analyze the cosmic radiations [13]. Another vital mission on space exploration is Radio Aurora Explorer (RAX) consisting of 3U CubeSats. RAX was launched for investigating the formation and distribution of natural ionospheric plasma turbulence [14], which is useful for enhancing the space weather forecasting and minimizes incurring damages to satellites and spacecraft technologies. Another remarkable CubeSat mission on space science is Mars Cube One (MarCO), which consists of 6U CubeSats. MarCO is the first flying CubeSat to deep space that was launched in 2018 by NASA to support relaying for telecommunications to Mars [15]. For completeness, all the aforementioned CubeSats missions are summarized in Table I; please also see references [4, 19, 20, 21, 22] for a complete list of academic and industrial CubeSats missions.

I-a Related Review Articles

Several review articles can be found in the literature of CubeSats, e.g., [19, 20, 21, 22, 23, 24, 25, 26, 27, 28], the summary of which can be found in Table II. For instance, a comprehensive survey on CubeSats missions was first presented in [19], which focuses on the history of different missions till 2009. The literature on these missions was further extended in [20] to the year 2013 where the statistics of CubeSats missions are briefly presented. The statistics in [20] include the number of missions launched, launch failures, operational missions, non-operational missions, and failed missions from the year 2000 to 2012. Reference [20] further provides the ratio of these missions in different parts of the world until 2012. In [21], a comprehensive survey about the space networks and interplanetary internet is provided. Reference [21] presents the concept of delay tolerant networking for deep space networks. However, the focus of [21] is not on CubeSats but towards networking aspect for deep space networks. In [22], Villela et. Al extended the statistic on CubeSats missions till 2018. They presented the statistics on the number of countries involved in CubeSat research, the success rate of CubeSat missions, and predicting that one thousand CubeSats will be launched in 2021. A brief survey on inter-satellite communication for CubeSats is presented in [23]. The focus of [23] is towards enabling inter-satellite communications by investigating the physical, network, and medium access control layers of the open system interconnection (OSI) model for small satellites.

Ref. Year Area of Focus
Bouwmeester et al [19] 2010 History of CubeSat missions from 2000 to 2009
Michael et al. [20] 2013 History and statistics of CubeSats missions from 2000 to 2012
Joyeeta et al. [21] 2013 Deep space networks and interplanetary internet
Thyrso et al. [22] 2016 History and statistics of CubeSats missions from 2000 to 2016
Radhika et al. [23] 2016 Inter-satellite communications for CubeSats
Scott et al. [24] 2016 NASA’s near earth network and space network for CubeSats
Yahya et al. [25] 2017 Antenna designing for CubeSats
Martin [26] 2018 History, statistics, and applications of CubeSats missions
Franco et al. [27] 2018 small satellite missions, antennas design, and networking
Anna et al. [28] 2018 Hardware challenges for CubeSat missions
This paper 2019 coverage and constellation issues, channel modeling, modulation and coding, networking, and future research challenges for CubeSats
TABLE II: Comparison of this paper with the existing survey articles.

In [24], the authors summarized the support, services, and future plans provided to the emerging CubeSat market by NASA’s Near Earth Network (NEN), Space Network (SN), and space communication and navigation network (SCaN). The work presented in [24] also discussed the capabilities of the NEN and SN, illustrating the maximum achievable data rates and data volumes for different orbit altitudes and slant ranges. The literature on the development of CubeSat antennas was summarized is [25], which discussed the antennas used for CubeSats such as horn, patch, dipole, reflector, and membrane antennas. Recently, the literature on evolution, constraints, policies, and applications of small satellites was presented in [26]. Davoli et. al presented an overview of the different aspects of small satellites, which include hardware components, antennas design, and networking [27]. Hardware challenges for CubeSat missions such as miniaturization, power control, and configurations are presented in [28].

In summary, most of the above surveys focus on details of missions, e.g., number of missions, number of CubeSats, launching dates, participating countries, and mission targets [19, 20, 26, 22]. However, only few surveys discuss the communications aspects of CubeSats, e.g., inter-satellite networking [21], antenna design [27], and delay tolerant networking [24].

I-B Contributions of this Paper

Despite the plethora of works on CubeSats, as highlighted in Section I-A, to the best of authors’ knowledge, there is no consolidated article which provides a comprehensive survey on the communication system aspects of CubeSats. More importantly, the existing surveys do not relate the effects of important technical considerations on CubeSats as communication systems, e.g., constellation design, cost constraints, etc.

In this paper, we provide an overview on the main properties of CubeSats that significantly influence the performance of the communication system. Then, we answer some research questions that can provide insights about CubeSat communications, e.g., example, how the constellation type, the CubeSat altitude, and the mission targets affect the link performance. We also address the question of designing efficient communication systems which account for these properties.

The paper first provides an overview on the constellation design for CubeSats. The choice of a particular constellation type depends on the goal of the mission. For example, if the target is to provide a global coverage for communication purposes, a larger number of satellites is required. We thereby discuss the required number of orbital planes and satellites to achieve this ubiquitous coverage. On the other hand, for remote sensing applications, fewer number of CubeSats are typically needed. Moreover, we also discuss the question of how to extend the data coverage using CubeSats for communications in rural and remote areas of the world, an uprising topic in CubeSats design.

Secondly, the impact of the satellite geometry on the communication channel model is discussed. We elucidate the evolution of the statistical channel models adopted for satellite communications from land mobile satellites (LMS) to CubeSats. Various channel models in the literature are compared according to their applicability to CubeSats.

Then, the link budget of CubeSats is investigated with respect to the satellite geometry, operating frequency, and channel modeling. The parameters of the link budget depend heavily on the characteristics of CubeSats, e.g., limited power and antenna gain. Next, we introduce the modulation and coding techniques usually adopted in CubeSats. Moreover, we show that how these schemes take into account the link budget and elevation angle towards the satellite so as to provide reliable communications. Also, we compare between these techniques and present the recommendations suggested by the consultative committee for space data systems (CCSDS).

Furthermore, the physical networking aspect of CubeSats is briefly discussed based on different communication technologies such as, radio frequency (RF) and free space optics (FSO), including Laser and visible light communication (VLC). We also discuss the routing protocols used for satellite-to-ground and inter-satellite links.

Lastly, we anticipate the future research directions and major open issues of CubeSats research, e.g., heterogeneous CubeSats-6G networks, software defined networking, Internet of space things (IoST), hybrid architectures, ubiquitous coverage, and machine learning.

The contributions of the paper are summarized as follows:

  1. We provide a comprehensive survey of CubeSats communications, which is envisioned to enable IoST.

  2. We discuss the most uprising topic in the domain, i.e., the extended coverage using CubeSats for communications in rural and remote areas of the world.

  3. We survey the technical aspects of CubeSats communications, including channel modeling, modulation and coding schemes, networking, constellation design, and coverage issues for CubeSats.

  4. We present several future research directions of CubeSats and its applications for earth remote sensing, space sensing, and global communications.

I-C Organization of the Paper

The article is organized as follows. Section II presents an overview of constellation designs, and coverage concepts for CubeSats. Section III covers channel modeling techniques, while Section IV discusses the link budget calculation for CubeSat communications. Sections V and VI explain modulation and coding methods, and networking techniques, respectively. In Section VII, we focus on future research challenges for CubeSat communications. Finally, Section VIII summarizes and concludes the survey.

Ii CubeSat Constellations and Coverage

Coverage of any CubeSat mission depends on different parameters such as number of satellites, number of orbital planes, elevation angle, inclination, altitude, orbital plane spacing, and eccentricity of the orbit. To date, CubeSats are deployed in the LEO orbits for lower cost and lower implementation complexity reasons. However, the footprint of LEO satellites is much smaller than the medium earth orbit (MEO) and geostationary earth orbit (GEO) satellites. To put this in perspective, more than 100 LEO satellites are required for global coverage, as compared to less than 10 MEO satellites [29].

Besides the altitude, constellation design is a major parameter to characterize the coverage of CubeSat missions [30]. We next present the three constellation designs for global coverage:

  • Walker constellations: Walker constellation design is symmetric, i.e., all the satellites have the same inclination and latitude. The parameters of Walker constellation are defined as inclination , number of satellites , number of equally spaced orbital planes , and relative phase difference between the planes . Based on these parameters, each orbital plane has number of satellites, whereas the inclination of all the planes is similar. The famous European GALILEO system uses Walker constellation design with three orbital planes at the inclination of with nine satellites in each plane, as illustrated in Fig. 1. Although Walker constellation design is simple to implement due to its symmetric nature, its coverage is constrained by the inclination angle. Latitudinal zones which are beyond the inclination angle of the orbital planes may not have any coverage.

    Fig. 1: Illustration of Walker constellation for GALILEO.

    To design and analyze a CubeSat constellation for a longitudinal global coverage, the minimum number of CubeSats per orbital plane, and the minimum number of planes required for a circular orbit can be determined as,

    (1)

    and

    (2)

    respectively, where is the ceil function and is the Earth central angle of coverage. By using the law of sine, the Earth central angle is obtained as follows [31],

    (3)

    where is the Earth’s radius, is the orbital altitude of the CubeSat, is the elevation angle, and is the slant range (see Fig. 2). The slant range can be determined by the law of cosines as,

    (4)

    To illustrate the effect of altitude and elevation angle on and , we plot (1) and (2) in Figs. 5 and 6, respectively. As shown in Figs. 5 and 6, the CubeSat requirement per orbital plane and the number of orbital planes increase with increasing the elevation angle and reducing the altitude. This is due to the direct relation between the Earth central angle , the elevation angle , and the altitude . Furthermore, for a fixed altitude, increasing the elevation angle from to , leads to increasing the number of planes and number of CubeSats per plane. However, reversing this scenario, i.e., keeping the elevation angle fixed and increasing the altitude from km to km, reduces the number of planes and the number of CubeSats per plane due to better coverage at higher altitudes.

    Fig. 2: Coverage geometry for CubeSats.
    Fig. 3: Illustration of polar inclined street-of-coverage constellation.
    Fig. 4: Illustration of Flower constellation in three different orbital planes.
    Fig. 5: Elevation angle versus number of orbital planes required.
    Fig. 6: Elevation angle versus number of CubeSats per plane.
  • Street-of-coverage constellations: These constellations consists of non-uniformly distributed polar inclined orbital planes. The separation between the orbital planes and their phase difference is designed in such a way that adjacent planes overlap with the coverage region, so as to provide global coverage. A major issue with these constellations is that the Earth coverage is not uniform, with the highest coverage at polar regions and poor coverage at equatorial regions, as shown in Fig.

    3. Missions which emphasize on coverage at the equatorial region require orbital planes to be equally spaced by ; however, such a constellation design requires longer deployment time and multiple launch sites [32].

  • Flower constellations: The idea of flower constellation was first introduced in 2003 [33]. Flower constellations consist of satellites following the same closed-loop trajectory in a rotating frame of reference [34]. The Earth-centered-Earth-fixed reference frame is used where all the satellites are synchronized and coordinated with the rotation of Earth. The orbital planes in Flower constellations satisfy the following condition [34]

    (5)

    where and are co-prime integers, is the time period of the rotating reference frame, and is the rational multiple of . Also, the semi-major axis, orbit inclination, perigee argument, and eccentricity of the orbits are the same. Furthermore, the mean anomaly of the -th satellite satisfies , where is the right ascension of the ascending node. Based on these conditions, the orbital planes are symmetric, and the satellites follow the same trajectory. This approach has two major issues: 1) the equivalency problem, since the mathematical expression leads to identical configurations, and 2) it is not clear that the original theory encompasses all possible configurations. Therefore, two-dimensional (2D) lattice theory was introduced in [35] by re-formulating the values of and . Although 2D lattice design leads to a fully symmetric constellation, it does not take into account the Earth oblateness effect. Therefore, inclinations of the orbits need to be adopted for elliptical orbits. To resolve this issue, three-dimensional (3D) lattice Flower constellations were introduced [36]. In 3D lattice constellation design, the argument of perigee affects the evenly distributed orbits on the same orbital plane. Fig. 4 shows an example of such constellation design, where a group of three orbits on the same orbital plane is used with the same inclination, eccentricity, and semi-major axis. Flower constellations provide some interesting orbital mechanics for flying formation and can support both regional and global area services.

The constellations design for satellites can also be configured for a specific mission. For example, in [37], Walker constellation was proposed for CubeSats to provide the air-traffic surveillance in Alaska region. Two perpendicular orbital planes were used with eight satellites in each plane, to provide 99 % coverage in Alaska. Also, the constellation of small satellites can be used to relay the data from the conventional satellites to the ground station. For example, in [38], nine CubeSats were placed in a near circular orbit at 496 km altitude to relay the information from the user satellite to the ground station. Under such scenario, the time for conventional satellite to ground was improved by 945 %.

Besides various constellation design, there is an on-going interest in swarm or cluster of small satellites based missions. The swarm of satellites can certainly improve the coverage of the missions both in space and on Earth. The concept of the swarm of satellites was introduced by U.S. Defense Advanced Research Projects Agency (DARPA) with the system of F6 (Future, Fast, Flexible, Fractionated, and Free-Flying) [39]. In F6 system, the traditional satellite was distributed among the cluster of sub-satellites where the resources among the sub-satellites were shared using inter-satellite communication. Although F6 system was the first step towards the swarm of small satellites, it was canceled after two attempts since an integrator was missing to pull the system together [40]. Inter-satellite communications and flight formation are the major concerns for the swarm of satellites [41]. Hence, efforts are made to optimize satellite-to-satellite coverage for the inter-satellite links. In [42], orbital parameters of the satellite were optimized to provide maximum coverage with six LEO satellites. Increase in the inter-satellite link improves the coverage of the mission. Danil et. al proposed a decentralized differential drag-based control approach for the cluster formation of 3U CubeSats [43]. In the absence of control strategy, the satellites in a cluster move apart from each other in the orbital plane. Therefore, it is important to model the aerodynamic drag force and reduce the relative drift between the satellites to zero. Recently, Cornell University launched 105 tiny size swarm of satellites, also called ChipSats in the KickSat-2 mission. The mission was successful where it was shown that forming a swarm of small free-flying satellites is possible [44]. These tiny satellites do not only reduce the cost further, but also improve the coverage in both space and on Earth. A freely-flying swarm of small satellites approach was proposed in [45] for providing connectivity to the IoT networks in the Arctic region. Three different orbit configurations of CubeSats were considered with three CubeSats and four ground stations. It was shown in [45] that freely-drifting swarm of CubeSats achieve overhead below 27 % and are, therefore, good candidates to support rural IoT networks.

Seamless global coverage can also be accomplished by proper beam coverage of the satellites. For instance, hybrid wide and spot beam schemes are presented in [46] for LEO satellites, where the wide beam is for large coverage area, and spot beams are for the high data access (see Fig. 7).

Fig. 7: Hybrid beam scheme for CubeSats in LEO.

Iii Channel Modeling

One of the significant issues for CubeSats missions is the lack of standardized communications channel model. Although the CCSDS provides some international standards, there are various issues to receive the telecommand signals in CubeSats using these standards. Mainly, these issues arise due to the error correction and detection codes used in CCSDS standards. Based on the CCSDS standards to be adapted for CubeSats communications, various channel models are proposed. Most of these models consider land mobile satellite (LMS) communication systems, where mobile devices on earth communicate with the CubeSats in LEO. In these models, the surrounding environment and atmospheric conditions are modeled using a statistical model.

In this section, we cover various well-known LMS channel models which can be used for CubeSats. The LMS channels can be broadly categorized into static and dynamic channel models. Static LMS channel models mainly consider the LoS direct path, LoS diffused path, and multipath [47]

. Multipath fading is modeled by using the well-known statistical distributions such as Rayleigh and Rice distributions, while the shadow fading is modeled using lognormal and Nakagami distributions. The dynamic models are based on Markov chains with different states for the LMS channel, where each state corresponds to the different propagation environments. Following are the major statistical channel models which can be used for CubeSats communications (also summarized in Table

III):

Loo’s Model [48]

It is one of the famous statistical channel models used for LMS systems. In Loo’s model, the amplitude of the line of sight (LoS) signal is modeled by using a log-normal probability density function (PDF), while the multi-path signals are modeled using a Rayleigh PDF [48]. Loo’s model assumes that the LoS component of signal, , undergoes a log-normal shadowing. The distribution of the signal envelope in Loo’s model is given by

(6)

where

is the variance of the multipath,

(7)

and are the mean and variance of the shadowing component, respectively, and is the zero-order Bessel function. This model has shown good agreement with the measured results in rural environment.

Corazza-Vatalaro’s Model [49]

Unlike Loo’s model, this statistical model combines Rician and log-normal distribution for the LoS signal, which is suitable for all different environments. The model is tested for both LEO and MEO earth satellites, where the theoretical results match the measured results. In this model, the PDF of signal envelope is the product of Rician and log-normal distributions, given by

(8)

where

(9)

is the Rice factor, and . Based on the values of , the above model can be reduced to any non-selective fading models. This model was extended in [50] by including the effects of phase variations in the shadowing and fading environment.

Saunders’ Model [51]

A geometrical approach is used to determine the blockage probability of the direct path. The geometry of the streets and buildings, which introduce a shadowing effect to the direct path, is taken into account. The probability that the direct path is blocked, i.e., when the height of the blocking building is larger than a certain threshold , can be written as

(10)

where is the variance of the building heights. By simple trigonometric relations, can be written as

(11)

where is the height of the receiver from the ground, is the distance between the building face and the receiver, is the width of the street, is the azimuth angle between the receiver and the satellite, and is the elevation angle.

Hwang’s Model[52]

This model extends the Corazza-Vatalaro’s model by including the independent shadowing which affects both the direct and diffused LoS link components. The PDF of the signal envelope in this model is given by

(12)

where

(13)

, are the independent log-normal distributions for the direct and diffused LoS links, respectively, is the LoS component of fading. Note that and are represented by (7) with parameters , , , and . When , , and then the fading component is absent while when and , this model tends to follow Corazza-Vatalaro’s model.

Ref. Year Multipath fading Shadow fading State Comments
[48] 1985 Rayleigh Lognormal Single Applicable only for rural environment, does not consider the Doppler effect usually present in LEO satellites
[49] 1994 Rice & Lognormal Lognormal Single

Applicable to both urban and rural environments because of the additional degree of freedom in modeling the LOS signal

[51] 1996 - Rayleigh Single Provide an insight about the deep fading probability in urban environment due to height of the surrounding buildings, width of the street, and the satellite geometry
[52] 1997 Rice & Lognormal with independent shadowing Lognormal Single Consider the Lognormal shadowing independent from the multipath fading, which allows more flexibility in fitting the real channel measurements compared to the Corrazz-Vatalaro’s model
[53] 1998 Rice Lognormal Single Account for the Doppler effect which makes it a better candidate for LEO satellites including CubeSats
[54] 2001 Rayleigh Lognormal Three Provide a Markov chain-based geometrical model which shows a good agreement to the real measurements over several frequency bands (been observed in all cases).
[55] 2003 Rice Nakagami Single Provide mathematically tractable model while fitting the real measurements for both narrow-band and wide-band systems
[56] 2006 Adaptive Adaptive Multi-state Provide a blind model, where the number of the Markov state and the distribution of the signal are not required a-priori
[57] 2014 Rayleigh Inverse Gaussian Single

Investigate experimentally the effect of tree shadowing and introduce the inverse Gaussian distribution to better model the shadowing

[58] 2019 Rayleigh (Adaptive) Nakagami (Adaptive) Multi-state Provide a Finite Markov chain-based model which is adaptive to the geometry of the CubeSats
TABLE III: List of statistical models for LMS channel

Patzold’s Model [53]

This model is similar to the Loo’s model; however, it also considers Doppler shift in the frequency Doppler asymmetrical power spectral density. These realistic assumptions increase the flexibility of the statistical model and fits well the measurements. The PDF of the signal envelope is given by

(14)

where is given in (7), , and is a parameter for the Doppler power spectral density, which can also control the fading rate. The derived PDF in [53] is a generalization of Rice density and is, therefore, more flexible. Also, this model has identical PDF expression to the Loo’s model; however, both models have different high order statistical properties, i.e., level crossing rates and average duration of fades.

Kourogiorgas’ Model [57]

This model investigates the first order statistics for the LMS channel in two different tree shadowing scenarios, i.e., intermediate and heavy tree shadowing, respectively. Small unmanned aerial vehicle (UAV) was used as a pseudo-satellite to experimentally investigate the effect of the tree shadowing. It was shown experimentally in [57] that Loo’s model offers the best accuracy among other models for the first order statistics of the received signal envelope. Furthermore, the authors also introduced inverse Gaussian (IG) distribution to model the tree shadowing [59]. Experimental tests were performed at a park, where the LoS signal was modeled as an IG distribution, while the multipath was modeled with a Rayleigh distribution. The PDF of signal envelope in [59] is given by

(15)

where is the PDF of the inverse Gaussian distribution given by

(16)

Here, and are the parameters of the IG distribution with variance .

Abdi’s Model [55]

This model characterizes the amplitude of the LoS signal by Nakagami distribution. This model is more flexible due to the closed form expressions of the channel statistics. The expression for the signal envelope in [55] is given by

(17)

where is the Nakagami parameter, is the spread parameter, and is the confluent hyper-geometric function [60]. This model fits well the Loo’s model and the measured results, for both narrow-band and wide-band systems. It has an additional advantage over previous models by having closed-form expressions for the channel statistics, e.g., the PDF, CDF, and moment generating function; therefore leading to a more tractable analysis.

Fontan’s Model [54]

This model considers a three-state Markov chain for three main propagation channel elements, i.e., the direct LoS, diffused LoS, and multipath signals. The states of the Markov chains are defined based on the degree of shadowing. This model is also tested for both narrow-band and wide-band conditions where the multipath delays are assumed to be exponentially distributed. The model was tested for L-band, S-band, Ka-band of frequencies in different environments with different elevation angles. Authors also provided a simulator that can generate time series of channel parameters such as Doppler spectra, phase variations, power delay profiles, and signal envelope.

Scalise’s Model [56]

This model is based on reversible jump Monte Carlo Markov chain (RJ-MCMC) to characterize the LMS channel. The first class statistical models work well under stationary conditions; however, they are not satisfactory when substantial changes occur to the propagation channel. Also, the multi-state Markov chain based models may not well characterize the real LMS channel. For example, in [54], different stats of the Markov chain represent the different channel elements, each having a fixed PDF. These assumptions make the model sensitive to the change in the propagation environment. Hence, RJ-MCMC model does not make any apriori assumptions on the propagation environment, the number of Markov states, and distribution of the envelope. This model was tested at Ku-band, where the model fits well with the measured results.

Nikolaidis’ Model [61]

This model uses a dual-polarized multiple input multiple output (MIMO) for LMS channel measurement. Channel capacities between 4.1-6.1 bits/second/Hz for both LoS and NLoS were calculated. The channel capacity varies significantly with the received signal pattern and elevation angle. Also, the channel capacities and correlation statistics were approximated using the stochastic channel models. Moreover, the mean quasi-stationary time of 41-66 seconds were found for different environments.

Salamanca’s Model [58]

Recently, finite state Markov channel with two sectors was introduced in [58] for LMS channel modeling. Its is an adaptive model that depends on the elevation angle of the satellite. More precisely, for low elevation angles, where the LOS signal is blocked, the fading amplitude is modeled by Rayleigh distribution. On the other hand, a Nakagami PDF can describe the distribution of the LOS signal envelope at higher elevation angles. The performance of the communication system over the proposed channel was simulated in terms of the bit error rate (BER) and throughput, following the CCSDS recommendations.

Iv Link Budget

Establishing a reliable communication link between a transmitter and a receiver is the ultimate goal of a radio link design in general. In particular, a CubeSat establishes two types of duplex radio links (uplink and downlink) with ground stations and with other CubeSats. Despite the prime role of the communication subsystem, the power that a CubeSat can dedicate is limited due to the weight and size constraints [62]. This section discusses the link budget expression for the downlink, i.e., CubeSat to ground commuications. The link design must ensure the ability to transmit and receive data directly from space to earth, or through one or more communication relays [63, 41].

A link budget is a set of parameters that define a communication link in terms of power for a reliable connection between the transmitter and the receiver. Based on the link budget, the signal to noise ratio (SNR) of the link is calculated, which measures the reliability of the link. The energy per bit to noise spectral density for the satellite to ground link (downlink) is given in [64] by

(18)

where is the transmitted power, and are the transmitter and receiver antenna gains, is the system temperature noise, is the free-space path loss, is the target data rate, and is the Boltzmann constant. The free-space path loss is given by

(19)

where is the distance between the ground station and the satellite and is the wavelength of the signal. Note that depends on the parameters of the LEO orbit such as the minimum elevation angle , angle between the position of CubeSat in orbit and the ground station , and CubeSat’s altitude from the center of Earth. Based on these parameters, is calculated as [65]

(20)

where is the Earth radius [66]. Fig. 8 depicts the relationship between these parameters and the distance.

Fig. 8: Schematic description of a LEO CubeSat trajectory
Fig. 9: Impact of the elevation angle on the distance between the satellite and the ground station.

For illustration purposes, we consider LEO orbits with three different altitudes and calculate the distance between the satellite and the ground station as shown in Fig. 9. It is clear from Fig. 9 that the distance between the ground station and the satellite is minimum when the elevation angle is degrees. The generalized SNR expression in (18) is valid for most of the CubeSat missions. However, the path loss varies for different missions due to the geographical location of ground stations, operating frequency band, different attenuation, and altitude of the orbits [67]. The frequency allocation for the CubeSat links, i.e., space-Earth, Earth-space, and inter-satellite, is regulated by the international entities for different applications. Typically, the frequency bands that are used for CubeSats are very high frequencies (VHF) or ultra high frequencies (UHF) amateur band [68]. However, some of the missions also used Ka-band [69], X-band [70, 71], S-band [72], L-band [68] and optical waves. To best characterize the effect of the elevation angle on the pathloss, we consider VHF-band and L-band frequencies and calculate the path loss with respect to the elevation angle as shown in Fig. 10. It is clear that the path loss is low at the degree elevation angle due to the shorter distance. Also, the path loss increases with the altitude of the satellite at higher frequencies. Further, in [31], a multi-band radio was proposed which covers a wide range of spectrum including microwaves, mm-waves, THz band, and optical waves for CubeSats. Link budget was calculated to show the effectiveness of the multi-band radios with continuous global coverage for the IoT networks. Table IV summarizes different frequency bands used for CubeSats [4].

Frequency band Frequency range
HF 20-30 MHz
VHF 145-148 MHz
UHF 400-402, 425, 435-438, 450-468, 900-915, and 980 MHz
L-Band 1-2 GHz
S-Band 2.2-3.4 GHz
C-Band 5.8 GHz
X-Band 8.2-10.5 GHz
Ku-Band 13-14 GHz
K-Band 18 GHz
Ka-Band 26.6 GHz
W-Band 75 GHz
Optical 400-700 THz
TABLE IV: Frequency bands for CubeSat missions

Also, CubeSat uplink frequency is kept higher than the downlink frequency to minimize the interference and attenuation, and to reduce the power consumption [64].

Fig. 10: Impace of the elevation angle on the path loss at different frequency bands.
Fig. 11: The energy per bit to noise spectral density for the downlink at the ground station vs the CubeSat altitude for various frequency bands and elevation angles.

V Modulation and Coding

A fundamental aspect for CubeSat communication systems is the design of the modulation and coding schemes. Since the weight and the cost of CubeSats are limited, there are high restrictions on the transmitted power. Hence, achieving a reliable communication with limited energy over land mobile fading channels is a challenging issue. The design of the modulation and coding scheme should take into account the proper trade-off between several parameters depending on the CubeSat mission. These aspects can be summarized as follows:

  1. the operational frequency band, e.g., UHF, S, X, and Ka bands, and the allocated bandwidth;

  2. the target operating rate;

  3. the duration of ground passes (i.e., the period at which the CubeSat is able to communicate with the ground station.)

For example, the available bandwidth at higher frequency bands like X-Band can reach  MHz, while the target bit rate is in the order of  Mbps for typical earth exploration CubeSat missions. Hence, binary modulation methods, along with low rate channel codes with high error-correction capabilities are preferable over higher-order modulation schemes with high rate forward error correction (FEC) codes. This is attributed to the reduction in the required power in the former case with the existence of more redundant data for efficient error correction, leading to higher power efficiency.

On the contrary, the available bandwidth at S-band for NASA missions is  MHz. Therefore, higher-order modulations, e.g., - phase shift keying (PSK) with rate- LDPC code, are essential to improve the spectrum efficiency [24, 73].

Another important aspect is the data volume needed to be communicated and the duration of ground passes. In fact, for some missions, the pass period is short while the amount of generated data is large. Hence, bandwidth-efficient communication systems with high data rates are required for reliable delivery of the information; thereby reducing the number of passes required.

Generally, choosing a suitable modulation technique for CubeSats requires a trade off between several metrics, e.g., the bandwidth and power efficiency, the BER performance, and the complexity of transceiver in the spacecraft. In the following, an overview of the most common modulation and coding schemes in CubeSats is presented.

Several modulation schemes are used in CubeSats such as quadrature phase shift keying (QPSK), offset QPSK (OQPSK), -PSK, - asymmetric PSK (PSK), - quadrature amplitude modulation (QAM) with . The performance of these schemes was investigated in [24] for various FEC coding rates and channel impairments like non-linearity. It was shown that higher-order modulations are vulnerable to the non-linear distortion, resulted from the power amplifier at the CubeSat, with an exception to -PSK, as it requires only a quasi-linear power amplifier.

The authors in [74, 75] suggested the use of Gaussian minimum shift keying (GMSK), where the payload of CubeSat is designed to cope with the link budget constrains for the system. It was shown through simulations that the CubeSat receiver can demodulate the signals with received power as low as  dBm [74]. This is attributed to the fact that GMSK signals have better spectral characteristics than OQPSK. Also, they have a constant envelope, allowing amplifiers to operate in the saturation region, which increases the power efficiency; however, GMSK has a poor error performance when compared to OQPSK[76]. In order to improve the BER, the authors in [76] proposed to employ a Viterbi decoder, leading to a higher computational complexity at the receiver.

OQPSK and rotated -QPSK were proposed as possible modulation techniques for CubeSats in [76]. This is attributed to the lower distortions resulting from the non-linearities of the amplifier, leading to better BER performance than classical QPSK. To improve the spectral characteristic of the OQPSK (i.e., reducing the out-of-band emissions), the CCSDS standard recommends using a filtered OQPSK scheme implemented using a linear phase modulator, i.e., OQPSK/PM. The OQPSK with phase modulator (OQPSK/PM) modulated signals have a constant envelope, permitting highly efficient nonlinear power amplification [77].

The effect of Doppler frequency shift on OQPSK/PM was investigated for a constellation of CubeSats around the Moon in [78]

. More precisely, the maximum Doppler and Doppler rate profile were estimated, and accordingly a frequency-tracking loop was designed to track the Doppler frequency and rate.

For higher spectrum efficiency, a hybrid modulation scheme where two parameters of the carrier are simultaneously modulated, e.g., the frequency and phase, can be used. For instance, a hybrid - frequency shift keying (FSK)/Differential QPSK modulation technique was presented in [79], leading to higher spectrum efficiency compared to -FSK. Another way to achieve highly-efficient CubeSat communication systems is by the joint design of higher-order modulation schemes with error-correcting codes, which is usually referred to as coded modulation framework, e.g., trellis coded modulation (TCM) [80]. One of these techniques is the bi-dimensional PSK-TCM which depends on -PSK modulations with the use of convolutional coding (CC) to introduce legitimate sequences between signal points joined by the trellis of the code.111Note that bi-dimensional refers to the in-phase and the quadrature components (). A generalization of this scheme involves several parallel modulating sequences (larger than ), which is referred to as -dimensional PSK-TCM. In this technique, the joint design of CC, PSK and multidimensional signal spaces provides a significant power gain, when compared to their sequential implementation, as shown in [81, Fig. B-4]. A comparison between TCM, CC, and turbo coding is conducted with OFDM signaling for LEO satellite channels within L-band and Ka-band in [82]. Turbo coded-OFDM achieves the lowest BER compared to CC-OFDM and TCM-OFDM systems.

The CubeSat link performance can significantly vary during the communication window due to environmental conditions (e.g., rain) or due to the change of the elevation angle for the observer ground station. For example, when the CubeSat rises from elevation up to , a variation of the up to  dB can be noticed, as shown in Fig. 11. Hence, adaptive modulation and coding schemes are required, as they can offer efficient communication over a wide range of signal-to-noise ratios by the proper interplay between the power and the spectrum efficiency. For adaptive modulation, the received signal strength should be calculated correctly at the satellite. In [83], a carrier to noise ratio, (), estimator was proposed based on fast Fourier transform.

To further increase the link performance, traditional multiple-input multiple-output (MIMO) schemes with multiple antennas in both the transmitter and receiver are usually deployed in terrestrial systems. However, most of CubeSats cannot support multiple antennas due to the size and cost limitations. Also, the separation between the antennas of the CubeSats would be too small to permit high performance gains, i.e., the channel is rank-deficient. Therefore, cooperative-communication techniques for CubeSats can be employed, where different spatially-distributed spacecrafts (each with single antenna) work together as a virtual entity. In this regard, a space-time based scheme (i.e., Alamouti’s code) was proposed in [84] for CubeSats to achieve high diversity gain. The BER performance of distributed MIMO and multiple-input single-output (MISO) schemes was simulated for various channel codes (e.g., convolutional, Reed-Solomon, LDPC, and turbo codes). It was found that combining distributed MIMO with channel coding leads to better error performance compared to single-input single-output (SISO) based schemes. On the other hand, several challenges face these techniques, including the phase synchronization between different satellites, and the induced latency in the inter-satellites links used for the cooperation between CubeSats.

Considering the complexity of the communication system, there are more restrictions on the computational complexity of the transceivers in CubeSats compared to those in ground stations. This is attributed to the limited power available for computations in the satellite due to the constraints on its size and cost. In this regard, a communication scheme between Lunar CubeSat network consisting of spacecrafts and an earth station was proposed in [85]. The system aims to achieve multiple access communication with a trade-off between the complexity at the CubeSat and the earth station. More precisely, the ground station uses uncoded CDMA in the uplink, allowing the possibility of having a simple decoder at the CubeSat. On the other hand, for the downlink the spacecraft employs the low-complexity sparse LDPC encoder followed by spread spectrum transmitter, leading to higher power efficiency at the expense of increasing the complexity at the earth station.

Frequency Band Mission type Modulation techniques
S-Band Space Research GMSK, filtered OQPSK
Earth Exploration GMSK, filtered OQPSK
X-Band Space Research GMSK, filtered OQPSK
Earth Exploration D -PSK TCM, GMSK, filtered OQPSK, -PSK, PSK with
Ka-Band Space Research  GMSK with precoding
Earth Exploration GMSK, filtered OQPSK, -PSK, PSK with
TABLE V: The CCSDS recommendations for modulation and coding schemes in LEO satellites.

Finally, we shed light on the CCSDS recommendations for modulation and coding schemes to be employed in the communication systems for satellites [81]. Table V shows the suggested modulation techniques for two types of missions (i.e., space research and earth exploration) operating at various frequency bands. It can be noticed that for S-Band, power-efficient modulation schemes with favorable spectral characteristics such as GMSK and filtered OQPSK are preferable. On the other hand, various higher-order modulation techniques with large spectrum efficiency are permitted at high-frequency bands, allowing the use of adaptive coding and modulation with fine granularity.

Vi Networking

Fig. 12: Networking architecture for CubeSats [86].

Networking for CubeSats can be classified mainly into two categories; CubeSat-to-Ground (C2G) networking and CubeSat-to-CubeSat (C2C) networking (see Fig. 12).

Vi-a CubeSat-to-Ground Communications

CubeSats utilize well-known UDP and IP protocols for C2G links achieving the data rate in few Mbps. For instance, DICE mission achieved the data rate of 3 Mbps operating in the UHF band [11]. In [87], the authors investigated the C2G link for Tianwang-1 mission which provided 125 Kbps of maximum data rate. In [88], VLC-based micro-satellites were used for space to earth links achieving the data rate of 9.6 kbps under perfect alignment. The altitude of the satellite in [88] was 400 Km with 40 Km footprint on Earth. A hybrid RF and optical communication approach is introduced in [89], where CubeSats are used as a relay satellite between the GEO satellites and ground station using both RF and optical links. However, pointing and acquisition are major problems for free space optical communications. Towards this direction, the first laser communication system was used in small optical transponder (SOTA) mission in 2014. SOTA was able to achieve the data rate of 10 Mbps for the downlink[90]. Lasers were also used in Aerocube OCSD mission for demonstration of C2G links providing high data rate and near-zero latency [91].

Fig. 13: Heterogeneous space information network.

Besides using various communication technologies for CubeSat links, networking issues are also investigated in the literature. For instance, Kak et. al introduced the use of software-defined networking (SDN) and network function virtualization (NFV) to provide end-to-end connectivity to the IoT networks [92]. SDN and NFV simplify the network management, improve network utilization, and provide fine-grained control for the system hardware [93]. Fig. 12 shows the network architecture of SDN for small satellites where the SDN controller is placed at the ground station (sink). The impact of different carrier frequencies and orbital parameters on the latency and throughput were investigated. Average end-to-end throughput of 489 and 35 Mbps was achieved for mmWaves and S-band, respectively [92]. A similar approach with fault recovery mechanism and mobility management using SDN was proposed in [94]. Also, the issue of controller placement in SDN based satellite networking was addressed where a three-layer hierarchical architecture was proposed. The domain controller, slave controller, and super controller were placed at the GEO satellites, LEO satellites, and ground station, respectively [95].

A hypothetical energy aware routing protocol by using contact graph routing (CGR) was proposed in [96]. CGR takes into account the prior information of contacts such as begin time, end time, and overall contact volumes for completing the path from source to the destination. In the CGR and extended CGR method, the CubeSats transmit to the ground station only when they have enough energy for data forwarding. Nevertheless, a Torrent-based approach is proposed in [97] for CubeSats to improve the downlink and uplink time for large files transmission. In CubeSat Torrent, large files are split into small chunks resulting in low latency. An analytical framework was proposed in [98], which formulates the data acquisition and delivering strategies for small satellites. It is shown numerically that joint optimization of data acquisition and data delivery can improve the delay-constrained throughput [98].

Vi-B CubeSat-to-CubeSat Communications

CubeSats can provide extended coverage in space and on Earth by working as an inter-satellite relay. However, coordination among the CubeSats requires C2C communications, which is a challenging task. The existing C2C link considers RF communications, highly directed lasers, and VLC. The latter two require accurate pointing among the CubeSats while the former is not suitable for high data rate applications and systems with sensitive onboard electronics. Most of the missions which employ C2C communications are based on either RF or lasers. The most ambitious mission which considers C2C links was QB-50, which consisted of a swarm of fifty CubeSats to study the upper thermosphere of Earth. Until now, a swarm of 36 CubeSats is launched which utilizes RF-based C2C communications. The use of TCP and UDP over IP was studied [99] for the C2C links in QB-50 mission. It was shown that TCP provides reliable C2C links while UDP has low latency. Recently, in [100], the authors investigated the QoS requirements for C2C link with massive MIMO. The results in [100] showed that using massive MIMO for C2C links improves the communication time; however, the use of massive MIMO increases the size of the CubeSats and the power consumption. A swarm of four CubeSats was developed in [101], which demonstrates multipoint to multipoint high data rates C2C links. S-band frequencies are used in [101] achieving 1 Mbps downlink, and 100 kbps crosslink data rates.

Besides RF and lasers based C2C links, recent advancements in free space optics and light emitting diodes (LED) technologies triggered the use of VLC for C2C links. The LED technology is advantageous over its counterparts for C2C links due to their low power consumption and lightweight. The feasibility of using LEDs for a hypothetical C2C link was examined in [102]. This work was mainly focused on minimizing the background illumination for hypothetical C2C links. In addition to the background illumination noise, the authors in [103, 104] investigated the effect of solar radiations on VLC-based C2C links. The solar radiations significantly reduce the SNR of the received signal. Using the transmit power of 4 Watts and digital pulse interval modulation, the data rate of 2 Mbps was achieved with BER = , for the transmission range of 500 meters. The proposed scheme is designed to comply with the limited size, mass, power, and cost requirements for the CubeSats. The feasibility of C2C links is numerically evaluated in [105] where it was shown that using higher frequencies reduces the communication time for C2C links. An inter-satellite scheme that employs channel coded CDMA has been proposed in [106]. The antennas are placed on the CubeSat faces and have been designed to maintain a fixed link between spacecrafts irrespective of their orientations, offering a  dBi gain.

Characteristics GEO LEO HAP UAV
Altitude 35,786 km 200-2000 km 17-25 km 0.1-20 km
Propagation delay 120-140 ms 1-15 ms Low Very low
Operational cost Low, GEO satellites do not require external power supply or physical maintenance Low, LEO satellites do not require external power supply or physical maintenance Recovery and redeployment required Changes with the payload type
Lifetime 10-15 years Few months or years depending on the mission Few days or weeks Depending on the power of battery, from few minutes to a single day
Deployment time Long, depending on the number of satellites in the constellation Long, depending on the number of satellites in the constellation Rapid deployment Very fast
TABLE VI: Comparison of various segments in space information networks.

In [107], the authors discussed the architecture for complex heterogeneous space information networks and provided protocols for sending the sensed data from the satellite to the ground in due time. Cooperative Earth sensing model was introduced, so as to replace the use of single satellites. Also, scheduling was briefly discussed, which is a complex issue for heterogeneous space information networks because of the different user’s demand, different types of CubeSats and C2C links, and uncertain space environment. Fig.  13 illustrates a heterogeneous complex space information network with conventional satellite networks, small satellites, and inter-networking between these networks. Table. VI compares the characteristics of major segments in space information networks, including GEO and LEO satellites, high altitude platforms (HAP), and UAVs. Recently, in [108], the authors discussed various applications that can be enabled by the space information networks; primarily focusing on providing connectivity to the IoT networks. Two application layer protocols constrained application (CoAP) and (Message Queuing Telemetry Transport) MQTT were analyzed for the space information networks where MQTT has larger variations in the good-put comparing to the CoAP protocol due to TCP bandwidth probing.

Vii Future Research Directions

CubeSats are envisioned to enable future wireless communications in space for various applications. Comparing to the existing satellite communication systems, CubeSats have significant features such as low cost and low altitude. However, the research on CubeSats for communications is still in the early phase and therefore has a wide variety of research problems. In this section, we point out the significant research challenges for CubeSats communications.

Vii-a Integration with Next Generation Wireless Systems

One interesting research area is the integration of CubeSats communications with next generation wireless networks such as 5G and beyond [109]. For example, Babich et. al have recently introduced an integrated architecture for nano-satellites and 5G [110]. The terrestrial communication link operates on mmWaves, while for the satellite to ground links and inter-satellite links, RF communication was used. For 5G and beyond systems, the intrinsic ubiquity and long coverage capabilities of CubeSats would make it a major candidate to overcome the digital divide problem in future wireless communications. As the existing studies are limited in this research area, investigating ways of integrating CubeSats with next generation networks, both at the physical and networking layers, is a promising research avenue for achieving the ambitious data metrics of 5G and beyond systems.

Vii-B Scheduling

Due to their small size, CubeSats have a limited number of onboard transceivers, which limits the number of communication contacts. Therefore, data scheduling is required to utilize the available transceivers efficiently. Hence, in [111], a finite-embedded-infinite two-level programming technique was proposed to schedule the data for CubeSats optimally. This technique considers stochastic data arrival and takes into account the joint consideration of battery management, buffer management, and contact selection. This framework demonstrated a significant gain in the downloaded data concerning the battery and storage capacities. Similarly, a scheduling algorithm was designed in [112] for CubeSats, which had four times better computational speed than using integer programming. The optimized scheduler takes into account the attitude control and orbital mechanics of the CubeSats to maximize their coverage. These theoretical scheduling models are required to be tested in a mission for validation. Also, these scheduling frameworks can be integrated with the UAVs and high altitude platforms (HAPs) to enable a more complete space information network.

From another perspective, a scheduling method for the tasks that should be executed by Cubesats is proposed in [113]. The scheduling algorithm opts for the number and the type of tasks to be executed in order to permit the solar panels of the energy harvesting system to operate close to their maximum power point, leading to higher power efficiency. The proposed scheduling algorithm can lead to about 5% decrease in the energy consumption with respect to systems without a task scheduler. As a future research direction, we propose investigating the problem of joint scheduling and data routing among multiple CubeSats, especially in cases which favor particular paths of CubeSats-to-ground communication routes.

Vii-C Software Defined Networking

The existing architectures of broadband satellite communication networks are inflexible due to its dependence mainly on the hardware. However, some of the recent works such as [114], [115], and [116] introduced the concept of using SDN for broadband satellite communication networks to improve their flexibility. Besides using it for broadband satellite communications, recently SDN was also proposed for CubeSat communications. For instance, SDN and network function virtualization (NFV) was used in [92] to provide connectivity to the IoT networks. It was shown in [93] that SDN and NFV improve the network utilization and control of the hardware, as well as simplify the network management. However, it was shown in [117] that the implementation of SDN/NFV for CubeSats communications faces multiple technical challenges. For instance, questions such as how the SDN protocols can be applied to the CubeSat gateways and remote terminals, how to perform dynamic network configurations to meet the QoS demand, and how to provide on-demand services without affecting regular operation of the network, remains open at the moment, and are promising open research topics.

Vii-D Towards Internet of Space Things

NASA aims at establishing a human colony on Mars by 2025, which will require connectivity beyond Earth [118]. To provide such intra-galactic connectivity, Internet of space things (IoST) is an enabling technology consisting of deep space CubeSats. Hence, there is a growing interest in the space industry to establish IoST networks, which is still in the early development phase. Besides, IoST will also provide extended coverage to the on-ground cyber-physical systems in rural areas. For instance, Iridium communications offer connectivity solutions for Earth remote sensing and space research with their 66 small satellites, also called SensorPODs [119]. CubeSats are going to play a significant role in the development of IoST networks where inter-satellite communication, in-space backhauling, and data forwarding are some of the exciting and challenging tasks.

Vii-E Hybrid Architecture

CubeSats can also be integrated with other communication technologies such as GEO and MEO satellites, HAPs, and UAVs. For example, in space information networks, CubeSats act as a relay between the GEO and MEO satellites. Similarly, CubeSats can perform back-hauling for HAPs and high altitude UAVs. There are some recent works such as [107], which discusses a hybrid architecture for CubeSats, conventional satellites, HAPs and UAVs. However, these architecture are only hypothetical models which require further validation. Also, the inter-linking between these entities of the space information network is challenging due to the dynamic nature of all these technologies and uncertain space environment.

Vii-F LoRa for CubeSats

The common Internet of Things (IoT) scenario involves connecting devices with limited energy over long ranges. In this regard, terrestrial-based low power wide area networks aim to offer low data rate communication capabilities over a wide area. An important communication technique for terrestrial IoT networks is the Low power long Range Wide Area Network (LoRaWAN). It is based on a novel frequency shift chirp spread spectrum modulation technique called LoRa[120, 121, 122]. Unfortunately, terrestrial-based LPWANs, including LoRaWAN, cannot offer ubiquitous coverage, especially for remote areas, e.g., desert, forests, and farms, mainly due to economic reasons. On the contrary, LEO nanosatellites for IoT can serve as a potential cost-efficient solution for this problem by providing global coverage. However, the modulation and multiple access techniques usually adopted in the terrestrial IoT systems can not be directly used in CubeSats, due to the Doppler effect and the propagation delay constraints.

In this regard, an architecture for an IoT based satellite system was provided in [123]. More precisely, a LEO Rosette constellation consisting of satellites in orbital planes with two additional polar satellites was proposed to assure global coverage. The compatibility of the communication protocols between the terrestrial IoT systems (e.g., LoRa and narrowband-IoT) and their satellite counterparts were also discussed. It was shown that cognitive radios mechanisms for interference mitigation or spread spectrum techniques should be used to permit the coexistence of both the terrestrial and satellite networks with acceptable interference level. Also, modifications on the existing higher layer protocols for IoT systems are required to decrease the overhead data to cope with the limited power and the delay involved in the satellite-based networks.

The feasibility of LoRa modulation in CubeSats, where the Doppler effect may have a non-negligible impact on the performance, was investigated in [124]. It was experimentally found that at higher orbits with altitudes more than km, LoRa modulation is immune to Doppler effect. On the contrary, the rapid variation in the Doppler frequency shift, e.g, when a lower altitude satellite flies directly above the ground station, leads to severe degradation in the performance, which reduces the duration of the radio communication session. From another perspective, the frequency shift spread spectrum in LoRa after some modification can be used to provide a multiple access technique as an alternative to direct sequence spread spectrum traditionally used in satellites [125]. It was shown that the BER performance of the proposed frequency and phase symmetry chirp spread spectrum are similar to the direct sequence multiple access.

Vii-G Machine Learning for Resource Allocation in CubeSats

It is fair to believe that more intelligence will be impeded in future wireless communication networks utilizing deep learning techniques. In CubeSat communications, one of the demanding challenges is the limited bandwidth resources which lead to low data rates, high latency, and eventual performance degradation. Therefore, it is envisioned to equip CubeSats with multi-band connectivity and smart capabilities to allocate power and spectrum resources in a dynamic approach across microwave, millimeter-wave, THz band, and optical frequencies

[126]

. This adaptive solution requires new transceivers and antenna systems, which are challenging research directions. Moreover, investigating and developing the performance of new resource allocation schemes will be the output of customized machine learning strategies. For example, a new multi-objective resource allocation scheme based on deep neural network (DNN) was proposed in

[126]

. Random hill-climbing algorithm was utilized instead of the back-propagation algorithm to adjust the weights of the neurons. This study was based on real satellite trajectory data of the Iridium NEXT small satellites, examining the influence of Doppler shift and heavy rain fade. The proposed DNN-based scheme resulted in improved multi-Gbps throughput for the inter-satellite links, and can be adopted in a multitude of future CubeSats communications problems of similar structures.

Viii Conclusions

CubeSats are envisioned to enable a wide range of applications such as Earth and space exploration, rural connectivity for the pervasive Internet of things (IoT) networks, and ubiquitous coverage. Currently, most of the research on CubeSats is focused on remote sensing applications. Unfortunately, little efforts are made to provide communication solutions using CubeSats, such as swarm of CubeSats for ubiquitous coverage, optical communication for high data rate, integration with future cellular systems for back-hauling, etc. Therefore, in this article, we review the literature on various communications perspectives of CubeSats, including channel modeling, modulation and coding, coverage, and networking. Also, some of the significant future research challenges are presented, which highlights how CubeSats technology is a key enabler for the emerging Internet of space things. Both the literature collection and the proposed research problems in this paper form a promising framework for addressing the world digital divide problem. In short, this paper can be a good starting point for the academia and industrial researchers focusing on providing communication solutions using CubeSats.

References

  • [1] “Why Facebook, SpaceX and dozens of others are battling over internet access from space,” http://fortune.com/2019/01/25/facebook-spacex-internet-access-space, accessed: 2019-06-20.
  • [2] K. Woellert, P. Ehrenfreund, A. J. Ricco, and H. Hertzfeld, “Cubesats: Cost-effective science and technology platforms for emerging and developing nations,” Adv. in Space Research, vol. 47, no. 4, pp. 663 – 684, 2011.
  • [3] R. N. Simons and K. Goverdhanam, “Applications of nano-satellites and cube-satellites in microwave and RF domain,” in 2015 IEEE MTT-S International Microwave Symposium, May 2015, pp. 1–4.
  • [4] E. Kulu, “Nanosatellite and cubesat database,” 2019. [Online]. Available: https://www.nanosats.eu/
  • [5] S. Bandyopadhyay, G. P. Subramanian, R. Foust, D. Morgan, S.-J. Chung, and F. Hadaegh, “A review of impending small satellite formation flying missions,” in 53rd AIAA Aerospace Sciences Meeting, 2015, pp. 1–17.
  • [6] H. Polat, J. Virgili-Llop, and M. Romano, “Survey, statistical analysis and classification of launched cubesat missions with emphasis on the attitude control method,” J. of Small Satellites, vol. 5, pp. 513–530, Jan. 2016.
  • [7] “World’s only connectivity solution designed for the poles,” https://www.keplercommunications.com/, accessed: 2019-06-20.
  • [8] “Radix,” https://space.skyrocket.de/doc_sdat/radix.htm, accessed: 2019-06-20.
  • [9] “Gomx-3,” https://gomspace.com/gomx-3.aspx, accessed: 2019-06-25.
  • [10] “Lemur-2,” https://space.skyrocket.de/doc_sdat/lemur-2.htm, note = Accessed: 2019-06-25.
  • [11] “Dice,” https://earth.esa.int/web/eoportal/satellite-missions/content/article/dice, accessed: 2019-06-25.
  • [12] “Quakesat,” https://space.skyrocket.de/doc_sdat/quakesat-1.htm, accessed: 2019-06-25.
  • [13] A. Budianu, A. Meijerink, and M. Bentum, “Swarm-to-earth communication in OLFAR,” Acta Astronautica, vol. 107, pp. 14 – 19, 2015.
  • [14] A. Poghosyan and A. Golkar, “Cubesat evolution: Analyzing cubesat capabilities for conducting science missions,” Progress in Aerospace Sciences, vol. 88, pp. 59 – 83, 2017.
  • [15] “Mars Cube One (MarCO),” https://www.jpl.nasa.gov/cubesat/missions/marco.php, accessed: 2019-06-25.
  • [16] “Integrated solar array & reflectarray antenna (ISARA),” https://www.jpl.nasa.gov/cubesat/missions/isara.php, accessed: 2019-06-25.
  • [17] “AeroCube-OCSD,” https://directory.eoportal.org/web/eoportal/satellite-missions/a/aerocube-ocsd, accessed: 2019-06-25.
  • [18] “Arcsecond space telescope enabling research in astrophysics (ASTERIA),” https://www.jpl.nasa.gov/cubesat/missions/asteria.php, accessed: 2019-06-25.
  • [19] J. Bouwmeester and J. Guo, “Survey of worldwide pico- and nanosatellite missions, distributions and subsystem technology,” Acta Astronautica, vol. 67, no. 7, pp. 854 – 862, 2010.
  • [20] M. Swartwout, “The first one hundred CubeSats: A statistical look,” J. of Small Satellites, vol. 2, no. 2, pp. 213–233, Mar. 2013.
  • [21] J. Mukherjee and B. Ramamurthy, “Communication technologies and architectures for space network and interplanetary internet,” IEEE Commun. Surveys Tuts., vol. 15, no. 2, pp. 881–897, Second 2013.
  • [22] T. Villela, C. A. Costa, A. M. Brandão, F. T. Bueno, and R. Leonardi, “Towards the thousandth CubeSat: A statistical overview,” Int. J. of Aerospace Engineering, vol. 5063145, no. 1, pp. 1–13, Jan. 2017.
  • [23] R. Radhakrishnan, W. W. Edmonson, F. Afghah, R. M. Rodriguez-Osorio, F. Pinto, and S. C. Burleigh, “Survey of inter-satellite communication for small satellite systems: Physical layer to network layer view,” IEEE Commun. Surveys Tuts., vol. 18, no. 4, pp. 2442–2473, Oct. 2016.
  • [24] S. Schaire, B. Horne, J. Schier, H. Shaw, G. Bussey, P. Celeste, Y. F. Wong, O. Kegege, Y. Zhang, C. Patel, D. Pierce, S. Altunc, D. Raphael, J. Burke, and L. V. Cooper, “NASA near earth network (NEN) and space network (SN) CubeSat communications,” in SpaceOps Conf., 2016, pp. 1–19.
  • [25] Y. Rahmat-Samii, V. Manohar, and J. M. Kovitz, “For satellites, think small, dream big: A review of recent antenna developments for CubeSats.” IEEE Antennas and Propagation Magazine, vol. 59, no. 2, pp. 22–30, Apr. 2017.
  • [26] M. N. Sweeting, “Modern small satellites-changing the economics of space,” Proc. of the IEEE, vol. 106, no. 3, pp. 343–361, Mar. 2018.
  • [27] F. Davoli, C. Kourogiorgas, M. Marchese, A. Panagopoulos, and F. Patrone, “Small satellites and CubeSats: Survey of structures, architectures, and protocols,” Int. J. of Satellite Commun. and Network., vol. 1, no. 17, pp. 1–17, 2018.
  • [28] A. Gregorio and F. Alimenti, “Cubesats for future science and internet of space: Challenges and opportunities,” in IEEE Int. Conf. on Electronics, Circuits and Sys., (ICECS), Dec. 2018, pp. 169–172.
  • [29] N. Lazreg and K. Besbes, “Design and architecture of pico-satellites network for earth coverage,” in Int. Conf. on Adv. Techno. for Signal and Image Process., (ATSIP), Mar. 2016, pp. 601–605.
  • [30] A. Marinan, A. Nicholas, and K. Cahoy, “Ad hoc CubeSat constellations: Secondary launch coverage and distribution,” in IEEE Aerospace Conference, Mar. 2013, pp. 1–15.
  • [31] I. F. Akyildiz, J. M. Jornet, and S. Nie, “A new CubeSat design with reconfigurable multi-band radios for dynamic spectrum satellite communication networks,” Ad Hoc Networks, vol. 86, pp. 166 – 178, 2019.
  • [32] Weilian Su, Jianwen Lin, and T. Ha, “Global communication coverage using cubesats,” in IEEE 7th Annual Computing and Commun. Works. and Conf., (CCWC), Jan 2017, pp. 1–7.
  • [33] L. McManus and H. Schaub, “Establishing a formation of small satellites in a lunar flower constellation,” The J. of the Astronautical Sci., vol. 63, no. 4, pp. 263–286, Dec 2016.
  • [34] M. E. Avendano and D. Mortari, “New insights on flower constellations theory,” IEEE Trans. Aerospace and Electronic Sys., vol. 48, no. 2, pp. 1018–1030, Apr. 2012.
  • [35] M. E. Avendaño, J. J. Davis, and D. Mortari, “The 2-D lattice theory of Flower constellations,” Celestial Mechanics and Dynamical Astronomy, vol. 116, no. 4, pp. 325–337, Aug 2013.
  • [36] J. J. Davis, M. E. Avendaño, and D. Mortari, “The 3-D lattice theory of Flower constellations,” Celestial Mechanics and Dynamical Astronomy, vol. 116, no. 4, pp. 339–356, Aug 2013.
  • [37] S. Nag, J. L. Rios, D. Gerhardt, and C. Pham, “CubeSat constellation design for air traffic monitoring,” Acta Astronautica, vol. 128, pp. 180–193, 2016.
  • [38] R. G. Perea-Tamayo, C. M. Fuchs, E. Ergetu, and L. BingXuan, “Design and evaluation of a low-cost CubeSat communication relay constellation,” in IEEE MTT-S Latin America Microwave Conf., (LAMC 2018), Dec. 2018, pp. 1–4.
  • [39] “DARPA moves ahead with fractionated-satellite system F6 program with solicitation for affordable satellite bus,” https://www.militaryaerospace.com/communications/article/16720493, accessed: 2019-07-03.
  • [40] “DARPAcancels formation-flying satellite demo,” https://spacenews.com/35375darpa-cancels-formation-flying-satellite-demo/, accessed: 2019-07-03.
  • [41] J. Alvarez and B. Walls, “Constellations, clusters, and communication technology: Expanding small satellite access to space,” in IEEE Aerospace Conf., Mar. 2016, pp. 1–11.
  • [42] I. Lluch and A. Golkar, “Satellite-to-satellite coverage optimization approach for opportunistic inter-satellite links,” in IEEE Aerospace Conf., 2014, pp. 1–13.
  • [43] D. Ivanov, U. Monakhova, and M. Ovchinnikov, “Nanosatellites swarm deployment using decentralized differential drag-based control with communicational constraints,” Acta Astronautica, vol. 159, pp. 646 – 657, 2019.
  • [44] “Swarm of 105 tiny sprite chipsats successfully deployed,” https://newatlas.com/sprite-chipsat-swarm-deployed/59994/, accessed: 2019-07-04.
  • [45] D. Palma and R. Birkeland, “Enabling the internet of arctic things with freely-drifting small-satellite swarms,” IEEE Access, vol. 6, pp. 71 435–71 443, Nov. 2018.
  • [46] Y. Su, Y. Liu, Y. Zhou, J. Yuan, H. Cao, and J. Shi, “Broadband LEO satellite communications: Architectures and key technologies,” IEEE Wireless Commun., vol. 26, no. 2, pp. 55–61, Apr. 2019.
  • [47] P. Arapoglou, E. T. Michailidis, A. D. Panagopoulos, A. G. Kanatas, and R. Prieto-Cerdeira, “The land mobile earth-space channel,” IEEE Veh. Techno. Mag., vol. 6, no. 2, pp. 44–53, Jun. 2011.
  • [48] Chun Loo, “A statistical model for a land mobile satellite link,” IEEE Trans. on Veh. Techno., vol. 34, no. 3, pp. 122–127, Aug. 1985.
  • [49] G. E. Corazza and F. Vatalaro, “A statistical model for land mobile satellite channels and its application to nongeostationary orbit systems,” IEEE Trans. on Veh. Techno., vol. 43, no. 3, pp. 738–742, Aug 1994.
  • [50] Sung-Chan Ko, Junghwan Kim, and Chen-Ying Yang, “Practical channel simulation model for the non-GEO land mobile satellite (lms) communications,” in IEEE 47th Veh. Techno. Conf., vol. 1, May 1997, pp. 411–415 vol.1.
  • [51] S. R. Saunders and B. G. Evans, “Physical model of shadowing probability for land mobile satellite propagation,” Electronics Lett., vol. 32, no. 17, pp. 1548–1549, Aug. 1996.
  • [52] Seung-Hoon Hwang, Ki-Jun Kim, Jae-Young Ahn, and Keum-Chan Whang, “A channel model for nongeostationary orbiting satellite system,” in IEEE 47th Veh. Techno. Conf., vol. 1, May 1997, pp. 41–45 vol.1.
  • [53] M. Patzold, Yingchun Li, and F. Laue, “A study of a land mobile satellite channel model with asymmetrical doppler power spectrum and lognormally distributed line-of-sight component,” IEEE Trans. Veh. Techno., vol. 47, no. 1, pp. 297–310, Feb. 1998.
  • [54] F. P. Fontan, M. Vazquez-Castro, C. E. Cabado, J. P. Garcia, and E. Kubista, “Statistical modeling of the LMS channel,” IEEE Trans. Veh. Techno., vol. 50, no. 6, pp. 1549–1567, Nov. 2001.
  • [55] A. Abdi, W. C. Lau, M. . Alouini, and M. Kaveh, “A new simple model for land mobile satellite channels: first- and second-order statistics,” IEEE Trans. Wireless Commun., vol. 2, no. 3, pp. 519–528, May 2003.
  • [56] S. Scalise, C. Alasseur, L. Husson, and H. Ernst, “SAT04-2: Accurate and novel modeling of the land mobile satellite channel using reversible jump Markov chain monte carlo technique,” in IEEE Globecom, Nov. 2006, pp. 1–6.
  • [57] C. Kourogiorgas, M. Kvicera, D. Skraparlis, T. Korinek, V. K. Sakarellos, A. D. Panagopoulos, and P. Pechac, “Modeling of first-order statistics of the MIMO dual polarized channel at 2 GHz for land mobile satellite systems under tree shadowing,” IEEE Trans. Antennas and Propag., vol. 62, no. 10, pp. 5410–5415, Oct. 2014.
  • [58] J. Lopez-Salamanca, L. O. Seman, M. D. Berejuck, and E. A. Bezerra, “Finite-state markov chains channel model for cubesats communication uplink,” IEEE Transactions on Aerospace and Electronic Systems, pp. 1–13, May 2019.
  • [59] C. Kourogiorgas, M. Kvicera, A. D. Panagopoulos, and P. Pechac, “Inverse Gaussian-based composite channel model and time series generator for land mobile satellite systems under tree shadowing,” IET Microwaves, Antennas Propagation, vol. 10, no. 6, pp. 612–616, Apr. 2016.
  • [60] M. Abramowitz, I. A. Stegun, and R. H. Romer, “Handbook of mathematical functions with formulas, graphs, and mathematical tables,” 1988.
  • [61] V. Nikolaidis, N. Moraitis, and A. G. Kanatas, “Dual-polarized narrowband MIMO LMS channel measurements in urban environments,” IEEE Trans. Antennas and Propag., vol. 65, no. 2, pp. 763–774, Feb. 2017.
  • [62] C. Benson, “Design options for small satellite communications,” in IEEE Aerospace Conf., March 2017, pp. 1–7.
  • [63] I. Latachi, M. Karim, A. Hanafi, T. Rachidi, A. Khalayoun, N. Assem, S. Dahbi, and S. Zouggar, “Link budget analysis for a LEO cubesat communication subsystem,” in Int. Conf. on Adv. Techno. for Signal and Image Process., (ATSIP), May 2017, pp. 1–6.
  • [64] O. Popescu, “Power budgets for cubesat radios to support ground communications and inter-satellite links,” IEEE Access, vol. 5, pp. 12 618–12 625, 2017.
  • [65] V. Almonacid and L. Franck, “Extending the coverage of the internet of things with low-cost nanosatellite networks,” Acta Astronautica, vol. 138, pp. 95 – 101, 2017.
  • [66] O. Popescu, J. S. Harris, and D. C. Popescu, “Designing the communication sub-system for nanosatellite CubeSat missions: Operational and implementation perspectives,” in SoutheastCon, Mar. 2016, pp. 1–5.
  • [67] I. Vertat, M. Pokorny, R. Linhart, and T. Kavalir, “Signal quality evaluation for picosatellite communication systems,” in Int. Conf. on Applied Electronics, Sep. 2012, pp. 331–334.
  • [68] D. Barbaric, J. Vukovic, and D. Babic, “Link budget analysis for a proposed Cubesat earth observation mission,” in Int. Convention on Info. and Commun. Techno., Electronics and Microelectronics (MIPRO), May 2018, pp. 133–138.
  • [69] O. Kegege, Y. F. Wong, and S. Altunc, Advances in Ka-Band Communication System for CubeSats and SmallSats.   NASA Goddard Space Flight Center; Greenbelt, MD United States, 2016.
  • [70] E. Peragin, F. Darnon, D. Belot, J.-P. Millerioux, J.-L. Issler, T. Dehaene, Y. Richard, G. Guillois, F. Sepot, and D. Simon, X BAND DOWNLINK FOR CUBESAT.   Utah State University, 2012.
  • [71] S. E. Palo, “High rate communications systems for CubeSats,” in IEEE MTT-S Int. Microwave Symp., May 2015, pp. 1–4.
  • [72] O. Ceylan, Y. Kurt, F. A. Tunc, H. B. Yagci, and A. R. Aslan, “Low cost S band communication system design for nano satellites,” in Proc. of 5th Int. Conf. on Recent Adv. in Space Techno., (RAST), Jun. 2011, pp. 767–770.
  • [73] Y. F. Wong, O. Kegege, S. H. Schaire, G. Bussey, S. Altunc, Y. Zhang, and P. Chitra, “An optimum space-to-ground communication concept for CubeSat platform utilizing NASA space network and near earth network,” in AIAA/USU Conf. on Small Satellite, 2016, pp. 1–19.
  • [74] A. Munir, N. Febrian, A. Daneraici Setiawan, and Chairunnisa, “Signal analysis of GMSK modulation-based cubesat automatic identification system receiver,” in Int.l Conf. on Info. Techno., Computer, and Electrical Engineering (ICITACEE), Oct. 2016, pp. 439–442.
  • [75] K. Pradhan, F. Pauline, G. Maeda, S. Kim, H. Masui, and M. Cho, “BIRDS-2: Multi-nation Cubesat constellation project for learning and capacity building,” in Proc. 32nd Annual AIAA/USU Conf. on Small Satellites, 2018.
  • [76] A. Gaysin, V. Fadeev, and M. Hennhofer, “Survey of modulation and coding schemes for application in CubeSat systems,” in Systems of Signal Synchro., Gener. and Process. in Telecommun., (SINKHROINFO), Jul. 2017, pp. 1–7.
  • [77] S. L. Identifiers, “Draft recommendation for space data system standards:radio frequency and modulation systems- part 1,” The Consultative Committee for Space Data Systems (CCSDS) 401.0-B-28.1, Tech. Rep., July 2018.
  • [78] A. Babuscia, D. Divsalar, K. Cheung, and C. Lee, “CDMA communication system performance for a constellation of CubeSats around the Moon,” in IEEE Aerospace Conf., Mar. 2016, pp. 1–15.
  • [79] I. Vertat, R. Linhart, L. Dudacek, and J. Mejzr, “Hybrid modulation as beacon replacement in CubeSat picosatellites,” in Int. Conf. Radioelektronika (RADIOELEKTRONIKA), Apr. 2013, pp. 297–301.
  • [80] G. Ungerboeck, “Trellis-coded modulation with redundant signal sets part i: Introduction,” IEEE Communications magazine, vol. 25, no. 2, pp. 5–11, 1987.
  • [81] “Bandwidth-efficient modulations: Summary of definition, implementation, and performance,” CCSDS 413.0-G-3: Green Book, Washington, DC, USA, International report, Issue 3, Feb. 2018.
  • [82] W. Fernando and R. Rajatheva, “Performance of turbo and trellis coded OFDM for LEO satellite channels in global mobile communications,” in ICC’98. 1998 IEEE International Conference on Communications. Conference Record. Affiliated with SUPERCOMM’98 (Cat. No. 98CH36220), vol. 1.   IEEE, 1998, pp. 412–416.
  • [83] I. Vertat, M. Pokorny, R. Linhart, and T. Kavalir, “Signal quality evaluation for picosatellite communication systems,” in In proc. Inter. Conf. on Applied Electronics, Sep. 2012, pp. 331–334.
  • [84] Z. S. Gibalina, V. A. Fadeev, K. A. Korsukova, M. Hennhöfer, and M. Haardt, “Estimation of capabilities of cooperative cubesat systems based on Alamouti transmission scheme,” in Systems of Signal Synchronization, Generating and Processing in Telecommun. (SYNCHROINFO), Jul. 2018, pp. 1–6.
  • [85] A. Babuscia, D. Divsalar, and K. Cheung, “CDMA communications systems with constant envelope modulation for CubeSats,” in IEEE Aerospace Conf., Mar. 2015, pp. 1–8.
  • [86] N. Saeed, M.-S. Alouini, and T. Y. Al-Naffouri, “To boldly go where no sensor has gone before: The movement to place IoT in radical new spaces,” arXiv preprint arXiv:1903.11996, 2019.
  • [87] S. Ennis and J. Dukes, “CubeSat networks: Balancing power with satellite-to-ground data throughput,” in IEEE Aerospace Conf., March 2018, pp. 1–18.
  • [88] A. Nakajima, N. Sako, M. Kamemura, Y. Wakayama, A. Fukuzawa, H. Sugiyama, and N. Okada, “ShindaiSat : A visible light communication experimental micro-satellite,” in Proc. Int. Conf. on Space Optical Sys. and App., (ICSOS), Oct. 2012, pp. 1–5.
  • [89] R. Welle, A. Utter, T. Rose, J. Fuller, K. Gates, B. Oakes, and S. Janson, “A CubeSat-based optical communication network for low earth orbit,” in AIAA/USU Conf. on Small Satellite, Mar. 2017, pp. 1–9.
  • [90] A. Carrasco-Casado, A. Biswas, R. Fields, and B. Grefenstette, “Optical communication on CubeSats enabling the next era in space science,” in IEEE Int. Conf. on Space Optical Sys. and Apps., (ICSOS), Mar. 2017, pp. 1–7.
  • [91] R. Welle, S. Janson, D. Rowen, and T. Rose, “CubeSat-scale high-speed laser downlinks,” in Proc. of the 13th Reinventing Space Conf., S. Hatton, Ed., 2018, pp. 7–17.
  • [92] A. Kak, E. Guven, U. E. Ergin, and I. F. Akyildiz, “Performance evaluation of SDN-based internet of space things,” in IEEE Globecom Workshops (GC Wkshps), Dec 2018, pp. 1–6.
  • [93] I. F. Akyildiz and A. Kak, “The internet of space things/CubeSats: A ubiquitous cyber-physical system for the connected world,” Computer Networks, vol. 150, pp. 134 – 149, 2019.
  • [94] S. Xu, X. Wang, and M. Huang, “Software-defined next-generation satellite networks: Architecture, challenges, and solutions,” IEEE Access, vol. 6, pp. 4027–4041, Jan. 2018.
  • [95] S. Xu, X. Wang, B. Gao, M. Zhang, and M. Huang, “Controller placement in software-defined satellite networks,” in 14th Int. Conf. on Mobile Ad-Hoc and Sensor Networks (MSN), Dec. 2018, pp. 146–151.
  • [96] M. Marchese and F. Patrone, “Energy-aware routing algorithm for DTN-nanosatellite networks,” in IEEE Global Commun. Conf., (GLOBECOM), Dec 2018, pp. 206–212.
  • [97] O. N. Challa and J. McNair, “CubeSat Torrent: Torrent like distributed communications for CubeSat satellite clusters,” in IEEE Military Commun. Conf., Oct. 2012, pp. 1–6.
  • [98] R. Liu, M. Sheng, K. Lui, X. Wang, Y. Wang, and D. Zhou, “An analytical framework for resource-limited small satellite networks,” IEEE Commun. Lett., vol. 20, no. 2, pp. 388–391, Feb. 2016.
  • [99] H. Bedon, C. Negron, J. Llantoy, C. Miguel Nieto, and C. O. Asma, “Preliminary internetworking simulation of the QB50 cubesat constellation,” in IEEE Latin-American Conf. on Commun., Sep. 2010, pp. 1–6.
  • [100] D. Bulanov, K. Ocheretyanyy, Y. Wang, and Q.-B. Chen, “Inter-satellite communication for LEO CubeSat network: QoS parameters and feasibility of massive MIMO,” J. of Commun. Techno. and Electronics, vol. 63, no. 10, pp. 1174–1182, Oct 2018.
  • [101] Z. Yoon, W. Frese, A. Bukmaier, and K. Brieß, “System design of an s-band network of distributed nanosatellites,” CEAS Space J., vol. 6, no. 1, pp. 61–71, Mar 2014.
  • [102] L. Wood, W. Ivancic, and K. Dörpelkus, “Using light-emitting diodes for intersatellite links,” in IEEE Aerospace Conf., Mar. 2010, pp. 1–6.
  • [103] D. N. Amanor, W. W. Edmonson, and F. Afghah, “Intersatellite communication system based on visible light,” IEEE Trans.on Aerospace and Electronic Systems, vol. 54, no. 6, pp. 2888–2899, Dec 2018.
  • [104] ——, “Link performance improvement via design variables optimization in LED-based VLC system for inter-satellite communication,” in IEEE Int. Conf. on Wireless for Space and Extreme Environments (WiSEE), Oct. 2017, pp. 7–12.
  • [105] D. Bulanova, Y. Wang, and Q. bin Chen, “Feasibility of intersatellite communication for CubeSats,” in Int. Conf. on Electronic and Info. Techno., (ICEIT), 2017, pp. 1–10.
  • [106] A. Budianu, T. J. W. Castro, A. Meijerink, and M. J. Bentum, “Inter-satellite links for cubesats,” in 2013 IEEE Aerospace Conference, March 2013, pp. 1–10.
  • [107] J. Du, C. Jiang, Q. Guo, M. Guizani, and Y. Ren, “Cooperative earth observation through complex space information networks,” IEEE Wireless Commun., vol. 23, no. 2, pp. 136–144, Apr. 2016.
  • [108] M. Bacco, L. Boero, P. Cassara, M. Colucci, A. Gotta, M. Marchese, and F. Patrone, “IoT applications and services in space information networks,” IEEE Wireless Commun., vol. 26, no. 2, pp. 31–37, Apr. 2019.
  • [109] F. Tariq, M. Khandaker, K.-K. Wong, M. Imran, M. Bennis, and M. Debbah, “A speculative study on 6G,” arXiv preprint arXiv:1902.06700, 2019.
  • [110] F. Babich, M. Comisso, A. Cuttin, M. Marchese, and F. PATRONE, “Nanosatellite-5G integration in the millimeter wave domain: A full top-down approach,” IEEE Transactions on Mobile Computing, pp. 1–10, 2019.
  • [111] D. Zhou, M. Sheng, J. Luo, R. Liu, J. Li, and Z. Han, “Collaborative data scheduling with joint forward and backward induction in small satellite networks,” IEEE Trans. Commun., vol. 67, no. 5, pp. 3443–3456, May 2019.
  • [112] S. Nag, A. S. Li, and J. H. Merrick, “Scheduling algorithms for rapid imaging using agile Cubesat constellations,” Advances in Space Research, vol. 61, no. 3, pp. 891 – 913, 2018.
  • [113] L. Slongo, S. Martínez, B. Eiterer, T. Pereira, E. Bezerra, and K. Paiva, “Energy-driven scheduling algorithm for nanosatellite energy harvesting maximization,” Acta Astronautica, vol. 147, pp. 141 – 151, 2018.
  • [114] T. Rossi, M. De Sanctis, E. Cianca, C. Fragale, M. Ruggieri, and H. Fenech, “Future space-based communications infrastructures based on high throughput satellites and software defined networking,” in IEEE Int. Symp. on Sys. Engineering (ISSE), Sep. 2015, pp. 332–337.
  • [115] R. Ferrús, H. Koumaras, O. Sallent, G. Agapiou, T. Rasheed, M.-A. Kourtis, C. Boustie, P. Gélard, and T. Ahmed, “SDN/NFV-enabled satellite communications networks: Opportunities, scenarios and challenges,” Physical Commun., vol. 18, pp. 95 – 112, 2016.
  • [116] E. Cianca, T. Rossi, M. Ruggieri, M. Presi, E. Ciaramella, L. Luini, C. Sacchi, G. Parca, and G. Codispoti, “Softwarization and virtualization as enablers for future EHF/FSO high throughput satellites,” in IEEE Global Commun. Conf., (GLOBECOM), Dec. 2018, pp. 1–6.
  • [117] “VITAL PROJECT H2020,” http://www.ict-vital.eu/, accessed: 2019-07-14.
  • [118] “Blast off: The internet of things is headed to Mars,” https://www.skyword.com/contentstandard/marketing/blast-off-the-internet-of-things-is-headed-to-mars/, accessed: 2019-07-27.
  • [119] O. P. Gupta, “Iridium NEXT SensorPODs: Global access for your scientific payloads,” in Proc. of the 25th Annual AIAA/USU Conference on Small Satellites, Aug. 2011, pp. 1–10.
  • [120] O. B. SELLER and N. Sornin, “Low power long range transmitter,” Feb. 2016, U.S. Patent 9,252,834.
  • [121] L. Alliance, “A technical overview of LoRa and LoRaWAN,” White Paper, November, 2015.
  • [122] M. Chiani and A. Elzanaty, “On the LoRa modulation for IoT: Waveform properties and spectral analysis,” IEEE Internet of Things Journal, pp. 1–1, May 2019.
  • [123] Z. Qu, G. Zhang, H. Cao, and J. Xie, “LEO satellite constellation for internet of things,” IEEE Access, vol. 5, pp. 18 391–18 401, Aug. 2017.
  • [124] A. Doroshkin, A. Zadorozhny, O. Kus, V. Prokopyev, and Y. Prokopyev, “Experimental study of LoRa modulation immunity to doppler effect in CubeSat radio communications,” IEEE Access, pp. 1–10, May 2019.
  • [125] Y. Qian, L. Ma, and X. Liang, “The performance of chirp signal used in LEO satellite Internet of Things,” IEEE Communications Letters, pp. 1–1, 2019.
  • [126] S. Nie, J. M. Jornet, and I. F. Akyildiz, “Deep-Learning-Based resource allocation for multi-band communications in CubeSat networks,” in IEEE Int. Conf. on Commun. Works., (ICC Workshops), May 2019, pp. 1–6.