Computer-vision integrated robotic arm for autonomous space operations

Arumugam J.P.; Aakaash A.; Sahoo R.K.; Camacho M.R. · Proceedings of the International Astronautical Congress, IAC (2025) · DOI: 10.52202/083083-0041

Resumen

Integration of Computer Vision (CV) into robotic weapons leads to revolution in human-robot partnerships in space exploration, providing autonomous operations into a complex and dynamic environment. This research presents a CV-based robot arm system designed for room assignments, which benefit from deep learning algorithm to detect real-time objects in microgravity, accurate manipulation and adaptive function. The system uses conventions and visual SLAM (simultaneous location and mapping) to increase spatial consciousness, causing the robotic arm to identify, track and manipulate elements such as equipment such as equipment, Payload and the spacecraft permission to interact with components. In addition, Bayesian decision network optimizes work execution under indeterminate conditions, enabling real -time adjustment of operating barriers such as light variation, speed disorders and unexpected environmental changes. A key feature of the system is its human-in-loop architecture, where astronauts can supervise or override the autonomous functions through the improved Augmented Reality (AR) interface. It increases additional ExtraVehicular Activities (EVAs), docking processes and efficiency in orbital repairs, reduces the workload and the risk of assignment for astronauts. Experimental results from the simulated microgravity environment show a 50% reduction in human intervention as compared to a 40% improvement in the completion of work and traditional teleport robot systems. In conclusion, the scalable, autonomous infrastructure mounting for planetary surface operations and the ability of AI-operated robotic weapons to enable deep space location based construction, marks a step toward permanent human-robot cooperation.

Instituciones: Universidad Nacional Mayor de San Marcos