SOUTHAMPTON, U.K. - AccelerComm has been selected to participate in a European Space Agency (ESA)-funded project developing a regenerative 5G non-terrestrial network (NTN) processor for Low Earth Orbit (LEO) satellite communications.
The Direct-To-Device 5G/6G Communication (D2SAT) project is part of ESA's Space for 5G/6G and Sustainable Connectivity program within the Advanced Research in Telecommunications Systems (ARTES) program. Antwerp Space leads the 26-month project, which is developing and validating a real-time regenerative digital processor to support 5G/6G NTN waveforms for direct satellite-to-device communications. The project is targeting Technology Readiness Level (TRL) 5.
Consortium effort
The consortium includes Antwerp Space in Belgium, Southampton-based AccelerComm in the U.K., Lasting Software in Romania, and Imec in Belgium. Antwerp Space is leading the project and will contribute system architecture, integration, low-power software-defined-radio technology, and regenerative-payload expertise. AccelerComm will provide physical-layer processing technology, while Lasting Software will contribute Layer 2 and Layer 3 software and satellite-channel emulation, and Imec will work on resource allocation and beam-management algorithms.
Related: Ericsson validates 6G building blocks in first pre-standard OTA session
The project will investigate a regenerative 5G NTN gNodeB, or base station, operating onboard a LEO satellite. Unlike a traditional bent-pipe architecture, in which a satellite primarily relays communications between users and ground infrastructure, a regenerative architecture performs additional communications processing onboard the spacecraft.
AccelerComm will contribute Layer 1, or physical-layer (PHY), processing for the demonstrator. These functions handle the encoding, modulation, and signal processing needed to transmit and receive data over a radio link. The planned architecture will use a field-programmable gate array (FPGA) to accelerate computationally intensive physical-layer functions, while a general-purpose processor will handle Layer 2 and Layer 3 communications protocols. AccelerComm's L1 technology is intended to support processing across FPGA, AI-engine, and CPU platforms.
The demonstrator will incorporate dynamic user allocation, beam hopping, beam steering and switching, and power-allocation strategies for LEO satellite networks. ESA also plans to model non-terrestrial-network conditions including fading, Doppler effects, and inter-beam interference.
ESA lists a 25 percent improvement in link budget and a 50 percent reduction in power consumption and physical size as development targets. The agency also estimates a 30 percent link-budget improvement from dynamic resource allocation and targets additional power and link-budget improvements through dynamic beam management.
The project will also implement advanced 3GPP NTN features and demonstrate digital beamforming in real time. The objective is to validate regenerative processing technologies for multi-beam, multi-user satellite communications under realistic and dynamic conditions.
The work reflects a shift toward placing more communications processing onboard LEO spacecraft, potentially increasing flexibility in how satellite computing, spectrum, power, and beam resources are allocated. At the same time, the approach places additional processing, power, thermal, and hardware-integration requirements on the spacecraft.
