Embedded AI for aerospatial navigation

ESL will contribute to the ClearSpace project by optimising the deep learning (DL) algorithms for the computation of the satellite’s relative position and altitude.

DL algorithms exert high computational demands, and can easily overwhelm the computational capabilities of a System-on-Chip (SoC). This project is confronted by a range of key challenges:

• the bandwidth available between the memory and the processing elements
• the limited energy budget of these embedded processing systems
• fluctuating temperatures between extremes of hot and cold

There are therefore two fundamental aspects to which ESL will contribute. Firstly, the design of the ultra-low power embedded SoC computing architecture, which will include ARM CPU processors and hardware accelerators mapped on field-programmable gate array (FPGA) platforms.

Secondly, ESL will optimise the initial DL code to maximise the use of this heterogeneous platform, comprising a number of processors and custom-made accelerators on FPGA.

Read more