Optimal control–based trajectory optimization for obstacle-avoiding robotic filament winding

With the advancement of modern industrial technology, square composite tubes have seen increasing application across various critical engineering fields. Traditional square steel and aluminum tubes are being progressively replaced by high-strength, corrosion-resistant composite alternatives. Robotic filament winding has become the dominant manufacturing approach due to its multi-degree-of-freedom (DOF) flexibility and high-precision control, which significantly enhance production efficiency and product quality. However, conventional obstacle-avoidance strategies based on path–velocity decomposition often fail to simultaneously satisfy dynamic feasibility and strict collision constraints, leading to suboptimal performance. To address this issue, this study proposes a unified optimal control framework for obstacle-avoiding trajectory planning in filament winding robots. The formulation integrates robot dynamics, joint constraints, collision-avoidance requirements, and time-optimal objectives within a pseudospectral transcription scheme. The resulting nonlinear programming problem is solved efficiently using automatic differentiation and numerical optimization techniques. Simulation studies conducted on a 6-DOF industrial robot demonstrate that the proposed approach generates dynamically feasible and smooth trajectories under constrained winding environments. Comparative evaluations against both a traditional industrial planning strategy (Rapidly-Exploring Random Trees (RRT)–Artificial Potential Fields (APF)–polynomial) and an optimization-based method, Covariant Hamiltonian Optimization for Motion Planning (CHOMP), indicate improvements in trajectory duration and jerk energy while maintaining strict dynamic and collision constraints.