Improved Whale Algorithm And Its Application For Autonomous Vehicle Path Plannin

In modern society with high intelligence,autonomous mobile robots are widely used in healthcare,catering,military,home and other fields.As a classical research model of autonomous mobile robots,the automated guided vehicle consists of sensing system,navigation system,control system and other parts.Among them,the navigation system is one of the most important parts,which can guide the automated guided vehicle to the target along a safe path.Path planning is the key technology in the navigation system,providing a safe path connecting the starting position and the target position for the automated guided vehicle.The research outcome on path planning is rich,but in the face of complex dynamic environments,traditional path planning methods suffer from high computational cost and low success rate,which are difficult to meet the requirements of practical applications,while meta-heuristic path planning methods are popular among researchers for their advantages of distributed computing,not relying on prior knowledge and strong robustness.The whale optimization algorithm(WOA)is a metaheuristic swarm intelligence optimization algorithm with the advantages of a simple structure and few parameters.The search agent update methods are similar to the optimization process of path nodes.Therefore,the WOA better matchs the model characteristics of the path planning problem.However,the original WOA suffers from premature loss of population diversity and premature convergence,which severely limits its optimization performance.To this end,this paper improves the basic WOA with the background of the path planning problem of the automated guided vehicle in complex dynamic environments and conducts research on the deep integration of the improved WOA with global path planning and local path planning.The specific work is as follows:(1)Modified covariance matrix adaptation evolution strategy(CMA-ES)based whale optimization algorithm with improved predation mechanism is proposed,namely MCIWOA.The algorithm first introduces the Brownian motion to improve the exploration ability.Then it adopts the improved predation mechanism to enhances the ability to escape from local minima.The prey group strategy is also used to enrich the diversity of the population.Finally,a improved local search mechanism is used to further improve the speed and accuracy of convergence.The performance of the proposed algorithm is verified by the CEC2013 benchmark suite and compared with other seven algorithms.The results show that the MCIWOA won the first place with the smallest average convergence error,showing strong optimization performance.(2)State changeable point and WOA-based node relocation algorithm namely SCP-WOANR is proposed.For the complex environment with more local minima and uneven distribution of obstacles,the state changeable point method(SCP)is proposed to avoid local minima and generate shorter initial feasible paths,and then the length and smoothness of the initial paths are optimized using the WOA-based node relocation method(WOANR).Finally,the performance of the proposed algorithm is verified by eight benchmark maps and compared with the other two algorithms.The results show that the SCP-WOANR has the smallest average path length,the highest average path smoothness and the smallest average planning time,showing better path search and optimization performance.(3)Whale optimization and dual evaluation strategy based dynamic window approach is proposed,namely WDESDWA.To address the local minima problem,a trajectory evaluation strategy based on the scanning point array is proposed.To address the dynamic obstacle avoidance problem,a trajectory evaluation strategy based on behavior matching is proposed.To address the shortcomings caused by fixed value evaluation function weights,such as poor adaptability and weak robustness,an adaptive weight mechanism based on MCIWOA is proposed.Finally,the performance of the proposed algorithm is verified in two static scenarios and four dynamic scenarios and compared with the other three algorithms.The results showed that the WDESDWA can avoid all obstacles in the six test scenarios and achieve the shortest planning time in three test scenarios.(4)Whale optimization and dynamic window based hybrid path planning algorithm is proposed,namely WDHPA.To address the problem of path planning security and optimality in complex dynamic environments,the WDHPA is proposed on the basis of the SCP-WOANR and the WDESDWA,which introduces a tracking term in the WDESDWA to track the global path generated by the SCP-WOANR and to avoid unknown obstacles during the movement.Finally,the performance of the proposed algorithm is verified by six complex dynamic scenarios,and the results show that the WDHPA can avoid all static and dynamic obstacles and follow a given global path to the target.In summary,this paper proposes the MCIWOA with strong optimization performance to address the inherent drawbacks of the original WOA and integrates the improved WOA with path planning to propose three path planning algorithms,namely SCP-WOANR,WDESDWA and WDHPA,respectively.Simulation results show that the proposed algorithms have high efficiency,strong robustness and adaptability,and can effectively solve the path planning problem of automated guided vehicle in complex dynamic environments.