Coordinated Control and Collision Avoidance for Flexible Dual-Arm Space Robots Using an IT2 FNN with Online PSO Tuning

This paper presents an adaptive control scheme for flexible dual-arm space robots, targeting robust coordination under strong dynamic coupling and multiple uncertainties. The controller features an interval type-2 fuzzy neural network (IT2 FNN) to compensate for lumped system uncertainties. To enhance autonomy, a dual-timescale learning framework is established: a particle swarm optimization (PSO) algorithm continually tunes the FNN antecedent parameters online, while a fast adaptive law updates the consequent weights. This is integrated with an artificial potential field (APF) based collision avoidance strategy, featuring a novel evaluation-suspension mechanism that decouples safety from performance optimization. Comprehensive simulations, including statistical analysis and ablation studies under a unified harsh environment, demonstrate that the proposed controller significantly improves trajectory tracking and vibration suppression while ensuring collision-free operation and maintaining computational feasibility.