Computer Vision-Based Joint Space Sensing and Communication Systems: Non-Source, Autonomy, and Low Latency

Abstract

The new space race is intensifying as thousands of satellites are poised for deployment into outer space each year. Companies are deploying low-earth orbit satellites at an unprecedented rate to create mega-constellations. Because outer space is becoming increasingly crowded and hazardous, establishing safe, efficient, and rapid inter-satellite communications is now a major challenge, essential for managing, maintaining, and inspecting vast space communication networks. Joint sensing and communication (JSC) presents a remarkable opportunity to enhance communication performance metrics by leveraging sensed surrounding information, including spectrum efficiency, beam training, and more. As such, applying JSC for space environmental sensing could significantly enhance the performance of inter-satellite communications in ever-crowded outer space. However, traditional radio frequency-based JSC faces significant challenges - such as latency and security concerns - when applied to outer-space satellite communications. This work proposes a novel approach to joint space sensing and communication-based on computer vision (CV) for target detection, tracking, and prediction to ensure effective satellite communications by integrating various communication and sensing techniques. We illustrate several network connection prototypes using appropriate computing management to enhance energy efficiency, latency, and covertness performance in various scenarios. Finally, we investigate and discuss the significant technological challenges and future research directions.