Validating Neuromorphic Autonomy for Deep-Space Operations

The increasing complexity and strategic importance of cislunar missions, from lunar gateways to in-situ resource utilization, demand a new level of onboard autonomy capable of operating for years without ground intervention. Conventional artificial intelligence systems are often ill-suited to this environment, constrained by high power consumption, vulnerability to temporal perturbations, and brittleness in the face of novel events. This paper presents a new class of bio-inspired, event-driven intelligence, instantiated on neuromorphic hardware, that provides a pathway to resilient autonomy. The efficacy of this approach is demonstrated through a series of high-fidelity, mission-relevant case studies, including a lunar rover surface mission and the long-duration monitoring of a cislunar gateway. The results show that the system can autonomously mitigate a ‘worst-case’ scenario involving simultaneous environmental faults, hardware degradation, and a direct cyber-attack, achieving a 92% mission success rate where conventional baselines underperform. This validated capability for onboard adaptive threat management is a critical enabler for the next generation of ambitious deep-space exploration.