Activity

Activities

gr-cssds

The gr-ccsds GNU Radio out-of-tree (OOT) module provides a set of transceivers implementing the CCSDS recommendations for telemetry and telecommand synchronization and channel coding.

The implementation follows the specifications desscribed in the CCSDS blue books CCSDS 131.0-B-3 , CCSDS 231.0-B-3 and CCSDS 401.0-B-28

Code Repository | Documentation

Radiation Evaluation

There is an increased interest in using commercially-grade SDR boards on board smaller satellites, sometimes with only some additional shielding. To achieve this, radiation test of commercial SDR boards is basic.
This activity will focus on the Characterization of Radiation resistance of the SDRs. The objective of this project is the first part of the categorization for 6 SDRs. For this reason we research the radiation each one of these boards can withstand according to the method and the materials used for its production according to previous research made and published.

Code Repository | Documentation

gr-leo

Gr-leo is a GNU Radio Out-of-Tree module that simulates the telecommunication channel between orbiting satellites and Ground Stations.
Gr-leo aspires to fill a gap with the implementation of a GNU Radio module that simulates the operation of an Earth-Satellite system and a variety of impairments that my occur during the space channel communication. For example, frequency shift due to the Doppler effect, the variable path loss due to the satellite's trajectory or the atmospheric effects along the path are proven to pose significant degradation on the communication channel.

Code Repository | Documentation

gr-soapy

gr-soapy is a GNURadio wrapper for the SoapySDR library.
The goal of this project was the implementation of gr-soapy, an Out Of Tree module for the GNURadio platform. The module provides source and sink blocks for a variety of SDR devices using SoapySDR, a generalized C/C++ library which provides abstraction in interfacing with different SDR devices and vendors.

Code Repository | Documentation

SDR Evaluation

The purpose of this activity is evaluate currently available SDR hardware and software from a satellite communications point of view. The evaluation will include both analytical assessments of the hardware and software capabilities, as well as lab tests supplemented by over the air usage of the systems.

Repository | Documentation

DVB-S2 LDPC SIMD

LDPC decoding is the most computational demanding task in a DVB-S2 receiver chain. This activity we aim at providing an optimized library capable of decoding LDPC frames in real-time on a modern general purpose processor. The development focus will be on leveraging modern hardware parallelism in terms of single instruction multiple data (SIMD), cache architectures, and multiple threading to reach high decoding efficiency. This task will be realized in the REDS institute of the HEIG-VD in Yverdon, Switzerland.

Repository

SDR DNN Inference deployment

Machine learning, and Deep Neural Networks (DNNs) in particular, are gaining traction in many fields of application obtaining very encouraging results. In the SDR domain there are relatively few attempts of adopting DNNs. One limiting factor to a broader adoption of these methods is the relative complexity of the deployment of trained models in an SDR context. In this activity we will design an out-of-tree GNU Radio module that allows easy and efficient deployment of inference DNN modules within a GNU Radio block diagram. This task will also explore the possibilities and the limitations inherent in the integration of GPU and FPGA acceleration in SDR context. This task will be realized in the REDS institute of the HEIG-VD in Yverdon, Switzerland.

Repository

Signal classification with deep learning

the objective of this proposal is to study and apply deep learning techniques in satellite signals, aiming to classify a previously detected signal (possibly using a DNN for the detection) in an a priori known family of signals, i.e., modulation schemes. An automated AI system that could recognize the received signal with increased accuracy and low latency is of high importance for an SDR system which aims at avoiding interference and/or changes the telecommunication scheme to establish communication with additional ground stations or other spacecraft.

Repository

Open Source FPGA toolchains

The main scope of this activity is to act as a foundation for FPGA related sub-activities by testing available open source tool chains for FPGA programming with a selection of popular FPGA platforms. Each tool chain will be evaluated against each FPGA platform in order to determine its strengths and weaknesses. During this process a basic set of FPGA programming tasks will be performed and evaluated. For toolchains that provide an API for extending hardware compatibility, support for additional hardware may be implemented depending on the subactivity timeframe and the maturity of the toolchain/API. As a result, a feature matrix will be created and possible investment on FPGA open source projects will also be assessed. All results and findings will be compiled in a sharable Technical Document.

Repository

Signal detection with deep learning

the objective of this proposal is to study and apply deep learning techniques in satellite signals, aiming to detect the presence of useful information in highly noisy time-series. An automated AI system that could discriminate between noise and useful payload with increased accuracy and low latency is of high importance for an SDR system which provides spectrum sensing functionality and aims at avoiding interference

Repository

SDR enabled spectrum monitoring demonstration

The primary objective of this activity is to provide a lightweight, yet complete spectrum analysis mechanism using the SDR technology. Initially the main focus will be on ground analysis. As a secondary objective, we will investigate optimizations and modifications that could enable the spectrum analysis to be performed in-orbit and especially by CPU/power limited devices. Such a system can be used around the world to identify possible misuses of the licenced spectrum, spot interference sources or identify spectrum opportunities in the unlicensed bands. In addition, spectrum monitoring can provide insights regarding the spectrum utilization and assist the selection of the band that will provide the performance for the satellite - earth communication. The main axes that should be followed are (a) on the minimization of the resources required for the spectrum analysis, (b) on the efficient multi-band monitoring, and (c) on the capability of interacting with A.I. systems for in-depth inspection. Furthermore, a study on the required arithmetic accuracy will be performed in order to be able to predict the performance in restricted processing units like FPGAs. Towards, the aforementioned paths a proper simulation environment will be implemented, with all the essential subsystems and parameters. The simulation will provide a solid software testbed, where everything will be evaluated before testing it on the real hardware.

Repository