Improved multi-objective coati optimization algorithm for trajectory planning of robotic manipulators

Trajectory planning is a core aspect of robotics, often involving coupled and conflicting performance objectives such as time, energy consumption, and trajectory smoothness. Traditional single-objective optimization approaches are insufficient to ensure efficiency, accuracy, and stability simultaneously. To address this challenge, this paper proposes an improved multi-objective coati optimization algorithm (IMOCOA) to enhance trajectory planning under high-dimensional and constrained conditions. First, a weighted objective function is constructed using the entropy weight method to achieve adaptive balancing among multiple objectives. Then, to overcome the limitations of the traditional multi-objective coati optimization algorithm—such as poor initial population diversity, premature convergence, and uneven distribution of solutions—tent mapping reverse learning, differential evolution, and evolutive Latin hypercube sampling initialization strategies are introduced. These enhancements improve population diversity and strengthen both global exploration and local exploitation capabilities. Under dynamic constraints, the proposed algorithm optimizes the control points of seventh-order nonuniform B-spline curves for multi-objective trajectory generation. Experiments conducted on the CoppeliaSim–UR10 platform validate the effectiveness of the proposed method. The results demonstrate smoother acceleration and jerk profiles, zero velocity and acceleration at start and stop points, and reductions in execution time, energy consumption, and jerk by 15.71%, 10.40%, and 5.97%, respectively. The results suggest that IMOCOA offers an efficient and reliable solution for complex multi-objective trajectory planning in robotic applications.