基于改进TSO_DWA算法的割草机局部避障OA

Aiming at the problems that the traditional Dynamic Window Approach(DWA)encountered difficulty in selec-ting the optimal path and generated non-smooth trajectories in areas with dense obstacles and regions with linearly moving obstacles,an improved DWA algorithm optimized by Tuna Swarm Optimization(TSO)foraging behavior was proposed for robotic lawn mower local path planning.Firstly,the initial positions of the tuna swarm were initialized using Fuch infinite folding chaos to enhance the search efficiency for optimal solutions.Its ergodicity could effectively avoid the problem of traditional random initialization falling into local optimum.Secondly,a learning rate ρ regulated the update step size of DWA's weight coefficients to strengthen path optimization,where perturbation terms r and coefficients σ were introduced to accelerate the convergence of optimal weights,and reduce the influence of the heading angle weight coefficient on the unsmooth path caused by the constant proportion of the heading angle weight coefficient in the complex environment.Fi-nally,the enhanced evaluation function scored candidate paths,and the optimal trajectory is determined by comparing iteration counts and evaluation scores.Simulation experiments and grassland experiments demonstrated that in the simula-tion environment,the TSO-DWA algorithm could plan a smoother and more reasonable motion path in the dense obstacle area and the linear moving obstacle area.In the grassland and pedestrian scenes,the mower had autonomous navigation ability,with the positioning error and the maximum tracking error≤0.16 m,which met the actual needs.