One of the principal problem of satellite communications are the atmospheric impairment effects on the downlink signal. At the frequency bands typically adopted for space missions (e.g., S, C and X band) such effects are mainly due to heavy precipitation (produced by convective rain events) and can be easily taken into account using climatological statistics of atmospheric attenuation for the linkbudget design. However, the frequency congestion occurring at the typically adopted bands is pushing the link frequency for satellite communications toward higher frequency bands. In the specific case of deep-space and interplanetary missions, where very large distances are involved, a high transmission capacity is needed, thus requiring higher link frequencies (such as the Ka band). This trend of moving towards high frequencies is supported by the need of smaller spacecraft-antenna sizes (for the same directivity) and of more directive ground-station antennas (for the same size). At such high frequencies, the atmospheric effects causing the signal degradation are not only convective rainfall events but also nonprecipitating clouds and moderate precipitation produced by stratiform rains. At Ka-band, clouds and rain attenuation can be of more than one order of magnitude larger than at X band. In this sense, classical link-budget design techniques are too conservative and, in some cases, turn out to be unreliable and new techniques are needed. Within this framework, we have already accomplished a feasibility study concerning the use of weather forecasts for the link-budget design of deep-space missions [1]. We have developed a model chain for the prediction of the radiometeorological factors that degrade the signal quality causing data-volume loss during a satellite-to-Earth transmission. On the basis of this rediction, the developed model chain allows the optimization of the operational parameters for the downlink transmission (e.g., the telemetry data-rate and the elevation
angle at which start and stop the transmission). The model chain is composed by a weather forecast, a radio propagation and a down-link budget module. The weather forecast module produces numerical weather predictions of the atmospheric state expected during the satellite transmission pass on a geographical grid-domain centered on the receiving ground-station. The atmospheric state is a vector describing the temporal evolution of thermodynamic variables (such as pressure, temperature, humidity, wind velocity and orientation) and microphysical variables (such as atmospheric particle concentration of water clouds and aerosol dispersions) in the 3-dimensional space. The second module converts the predicted atmospheric state vector into radiopropagation variables (i.e., atmospheric path attenuation and rightness temperature). Such conversion is accomplished through a satellite data simulator that produces radiopropagation variables as measured by a groundbased
antenna (given the link frequency and the antenna elevation angle) taking into account the gas absorption and the microphysical properties of cloud droplets and hydrometeors. Finally, the last module exploits the predicted radiopropagation variables for the optimization of the downlink and the maximization of the transferred data-volume. The output of the last module are the operational parameters (i.e., the transmission optimal-configuration for the predicted atmospheric state during the pass) and the expected transferred data-volume and signal-to-noise ratio. As previously stated, this model-chain was developed within a feasibility study applied to the BepiColombo ESA mission to Mercury (launched in October 2018) considering the two ESA receiving ground stations (Cebreros and Malargüe). Results, evaluated with respect to classical link-budget design techniques, highlighted a potential gain of 20% to 25% on the total yearly received data volume when using the developed model chain for the link budget design at Ka band [2]. In this work we will show some preliminary results of the application of the developed prediction model-chain to the Hayabusa-2 JAXA mission to the asteroid 162173 Ryugu (that was launched in 2014 and reached the asteroid in June 2018). Thanks to the availability of weather-forecast simulations on a geographical area centered on the receiving ground-station, for the prediction of the optimal datarate and signal-to-noise ratio we will apply a stochastic approach involving the simulated geographical grid-domain. The model-chain will be validated through the comparison of the predicted signal-to-noise ratio with the one measured by the Hayabusa-2 TTCP during the pass.

The steadily increasing data rates from new instrument developments onboard Earth Observation satellites calls for higher downlink data rates, but also for new concepts to process this data onboard efficiently to optimise the data downlink process for reduced latency of data reception.
Another approach is using constellations with high degree of interconnections between spacecrafts for routing the collected data through a virtual network up to a ground station. In this context, a new trend is rising which is to consider a spacecraft within a constellation as a data processing node allowing a user to execute an onboard application and retrieving only the relevant information. That point is particularly challenging, as at the time of manufacturing of the spacecraft the application is unknown.
Due to the orbital constraints in LEO, the use of relay satellites has been implemented. The SpaceDataHighway (based on EDRS infrastructure) is offering such a service to repatriate user data via its GEO-based satellites whenever the optical EO satellite is in eclipse. User data can currently be repatriated with 1.8 Gbps, evolving to 3.6 Gbps for future GEO data relay nodes currently under development. This technology also calls for an optimized on-board data chain with high flexibility.
The presented high-speed data chain comprises several essential elements: Instruments or payloads that provide the data, one or more payload processors which process the data from an instrument in a way specific to that instrument, a data storage unit for storing the data until it can be transmitted to ground, a data compression processor for reducing the volume of the data, saving room in the data storage unit and reducing the downlink data-rate requirements, an RF and optical downlink and an on-board network for connecting these elements together.
The elements forming the high-speed data chain and the concepts behind are shortly introduced in the following:
On-Board SpaceFibre Network
The concept for the on-board network is to use the emerging, high performance, high reliability, high availability and extremely versatile SpaceFibre standard to interconnect all of the data chain elements, providing consistent interfaces and common configuration, control and housekeeping capabilities as well as very high data rates.
Payload Processing
The concept for the payload processing is to combine a mix of powerful processing elements (programmable DSP processors, systolic arrays and reconfigurable FPGAs) interconnected with SpaceFibre to provide a flexible processing resource.
Compression Unit
The concept for the data compression is to contribute to and use the emerging CCSDS standards implemented in a powerful reconfigurable FPGA using SpaceFibre interfaces to connect to the on-board network to provide a modular data compression system.
Data Protection System
The concept for data protection is to build on a patented file-based data protection mechanism that operates at the file level, has a relatively low implementation cost, and which provides protection against the large data drop-outs expected in high rate downlinks.
Solid State Mass Memory
The concept for the data storage unit is to design a very flexible system which interconnects IO, memory modules and data processing functions using an internal SpaceFibre network and which is adaptable to suit a broad range of input data rates and storage capacities as required by different missions.
Radio Frequency Data Link
The concept for the RF data link is to improve the symbol rate and modulation method used in previous RF data links to provide significantly higher data rates (5 Gbps) and to use dual Polarization to implement two channels with an overall data rate of 10 Gbps.
Optical Data Link
The concept for the optical data link is to build on previous successful optical data link work at DLR and to design a third-generation system which is compatible with the emerging CCSDS standard for optical downlinks and which can support a net data rate of 8 Gbps. The design will also support substantially higher data rates using DWDM techniques.
Concerning the employed technologies, there is a move toward in orbit re-configurability which means that either software or reprogrammable architecture shall be used. This is clearly in favour of FPGA and dismisses the use of ASIC’s except for some generic functions. System on Chip (SoC) become available, mixing FPGA fabric with processors (either hard or soft core) on the same component. Flash components have been proven in flight, leading to efficient big memories at reasonable cost and with a very limited footprint.
Data streams are very diverse in terms of contents and throughput making it more and more necessary to define a standard able to support on the same medium the concurrent exchange of user data, control and command messages at very high data rates. In case of a point to point architecture a router will be also necessary.
The paper will provide details on the concepts, their implementation as well as employed technologies.