Neuromorphic Computing: A Bio-Inspired Approach to High-Altitude Platform Stabilization with Visual Odometry

Space-based navigation systems face critical challenges requiring innovative approaches beyond conventional visual navigation methods for dynamic environments and power-constrained platforms. This paper presents a visual odometry framework for neuromorphic sensors that addresses these fundamental limitations through bio-inspired event-based sensing technology. Traditional frame-based navigation systems are hindered by high power demands, computational complexity, and motion blur which severely limit their effectiveness for space-based Position, Navigation, and Timing (PNT) applications. Neuromorphic sensors address these fundamental limitations with their high dynamic range and motion-blur resistance. To leverage their strengths, we developed a purely event-based visual odometry algorithm that processes asynchronous event streams to compute five-degree-of-freedom motion parameters in near-space conditions. The algorithm's architecture employs a motion optimisation framework utilising motion compensation and contrast maximisation techniques optimised for sparse temporal data. Testing on real flight data demonstrated robust performance under extreme conditions, with the algorithm successfully converging for 82.47% of time windows despite complex, multi-axis motion dynamics. The motion compensation produced visually sharp, stabilised event representations, with contrast metric improvements spanning two orders of magnitude. These findings validate neuromorphic computing as a transformative technology for next-generation space-based PNT systems, particularly for emerging LEO constellations where power efficiency is paramount. The approach enables new capabilities for spacecraft relative positioning and autonomous navigation in challenging lunar and Earth orbit applications. By successfully translating bio-inspired visual processing principles to space navigation, our work advances the development of more efficient, responsive and reliable navigation technologies essential for future space missions.