FPGA-Based Neural Network Accelerators for Space Applications: A Survey

The survey discusses the use of FPGA-based neural network accelerators in space applications, emphasizing their advantages in flexibility, cost, and radiation tolerance. It reviews existing literature, identifies trends and gaps, and suggests future research directions to improve onboard computing systems for tasks like autonomous operations and sensor data analysis. This work is significant for practitioners looking to leverage FPGAs for high-performance computing in space missions.