On-board event-based debris detection and collision avoidance for leo satellites expro plus

Objective: To develop an on-board system for autonomous, real-time debris detection, tracking, and avoidance using amulti-sensory payload.Description:The growing presence of space debris in Low Earth Orbit (LEO) threatens satellite operations and spacesustainability. Small, fast-moving debris is difficult to detect yet capable of inflicting significant damage. Current avoidance strategies heavily depend on ground-based tracking and human-in-the-loop decisions, which can be slow and constrained by communication delays. This proposed activity builds upon a precursory ESAactivity NEU4SST - Neuromorphic Processing for Space Surveillance and Tracking, which produced valuable insights and a limited dataset. This dataset however was generated with a ground-based sensing focus and lacks realism, metadata richness, and orbital context required for space-borne applications.The aim of this activity is to demonstrate autonomous collision avoidance capabilities through real-time spacedebris detection and avoidance in LEO using event-based datafusion. To achieve this, this activity will develop a realistic, high-fidelity simulated dataset that mimics the operational environment of a spacecraft equipped with an event-based camera and navigation sensors (GNSS, IMU). This includes dynamic scenes with sub-10 cm debris, representative lighting conditions, orbital motion, and synchronized multi-sensor outputs. This dataset will be used tovalidate real-time detection, tracking, and collision avoidance algorithms based on Machine Learning. Edge-based processing algorithms will be demonstrated running on representative space grade or space-relevant embedded hardware platforms to evaluate their performance and feasibility under hardware constraints. In addition to the simulation-based prototype, the activity will develop a testbed prototype that integrates a physical event-based camera with a processing unit capable of executing the developed algorithms. Thissetup will allow demonstration of key capabilities, including object detection, data fusion, andautonomous decision-making.Thisactivity encompasses the following tasks:- Design and implement an on-board autonomy architecture using an event-based optical camera to enable decentralized decision-making for collision avoidance across spacecraft.- Develop edge-based processing algorithms for real-time debris detection and tracking, leveraging simulatedand semi-simulated datasets from the on-board sensory suite to estimate debris size, assess conjunction risk,trigger alerts and to optimize autonomous manoeuvres.- Implement on-board autonomous conjunction risk assessment and collision avoidance manoeuvre planning,including demonstration of the system on low-power space-grade/space-relevant hardware to evaluateperformance and reliability in constrained environments.- Simulate and validate high-fidelity LEO scenarios in controlled lab environments, including variable debrisdensities, orbital regimes, and dynamic interaction scenarios,tobenchmark system performance underrealistic operational conditions.Deliverables: Prototype, ReportsThis activity is currently in the GSTP Work Plan but will only be implemented after confirmation of financial support from Delegations. The final list of participating states will be known at tender issue.