By submitting this form, you are agreeing to the Terms of Use and Privacy Policy.
With varying degrees of effectiveness, Field Programmable Gate Array (FPGA) devices have been utilised in space for more than ten years.
Due to their vulnerability to involuntary reconfiguration brought on by Single Event Upsets (SEU) brought on by radiation, few reprogrammable electronics have up until this point been utilised aboard European spacecraft.
The fundamental benefit of an FPGA over a discrete circuit or Application Specific IC (ASIC) counterpart is the simplicity with which its functionality may be altered once a product has been created. Additionally, FPGA can be more energy efficient than a comparable discrete device and use less board area.
The Global Space grade FPGA market accounted for $XX Billion in 2023 and is anticipated to reach $XX Billion by 2030, registering a CAGR of XX% from 2024 to 2030.
Xilinx launches a new reconfigurable space-grade chip optimised for local machine learning on orbit. A new processor for in-space and satellite applications has been created by space-specific semiconductor manufacturer Xilinx, and it sets a number of firsts: It is the first process on a 20nm node that is approved for usage in space, delivering advantages in terms of power and efficiency.
Additionally, it is the first to specifically allow fast machine learning using neural network-based inference acceleration.
The processor is a field programmable gate array (FPGA), users may change the hardware to meet their individual requirements because the chip is essentially user-configurable hardware.
For a few reasons, Xilinx’s new chip has a lot of potential for the satellite market: it’s a major leap in processing size, since the company’s previous conventional tolerant silicon was only available in a 65nm standard.
That means significant improvements in terms of size, weight, and power efficiency, all of which translate to significant savings when it comes to in-space applications, because satellites are designed to be as lightweight and compact as possible to help defray launch costs and in-space propellant needs, both of which represent significant expenses in their operation.
Finally, due to its reconfigurable nature, on-orbit assets may be reprogrammed instantly to undertake various tasks, including the execution of local machine learning algorithms.
Therefore, it is theoretically possible to change one of them in an Earth observation satellite from handling tasks like tracking cloud density and weather patterns to drawing conclusions regarding, for example, deforestation or strip mining.
That adds a tonne of freedom for satellite constellation operators wanting to relocate where the market most urgently requires it.