Buffer Compliance Control of Space Robots Capturing a Non-Cooperative Spacecraft Based on Reinforcement Learning

Aiming at addressing the problem that joints are easily destroyed by impact torque during process of space robot on-orbit capturing a non-cooperative spacecraft, reinforcement learning control algorithm combined with compliant mechanism is proposed to achieve buffer compliance control. The can not only absorb energy through deformation its internal spring, but also limit safe range combining strategy. First all, dynamic models and target spacecraft before capture obtained using Lagrange approach Newton-Euler method. After that, based on law conservation momentum, constraints kinematics velocity, integrated model post-capture hybrid system derived. Considering unstable system, for stable associative search network employed approximate unknown nonlinear functions, an adaptive critic utilized construct signal tune network. numerical simulation shows scheme reduce acting 76.6% maximum 58.7% minimum in operation phase. And phase, were limited within safety threshold, which avoid overload damage joint actuators.