Research On Optimizing Path Planning Model By Integrating Improved RRT* Algorithm And ACO Algorith

Path planning is an essential component of an Autonomous Underwater Vehicle(AUV)exploration system.The path planning for an AUV should take into account optimization objectives,path environment,terrain and obstacles,as well as constraints related to the robot’s own performance.Various algorithms can be utilized to solve this path planning problem,among which the RRT* algorithm and the ACO algorithm have demonstrated relatively good performance.However,the marine environment is characterized by complex and diverse terrain with frequent variations,a higher probability of encountering new obstacles,and the commonly used path optimization algorithms suffer from issues such as long sampling time,high computational complexity,excessively random path planning results,susceptibility to local optima,and slow convergence speed.To address the challenges associated with the RRT* algorithm,including long sampling time,high computational complexity,and excessively random path planning results,a multistrategy improved algorithm called NTDRRT* is proposed.Firstly,a node attraction strategy is introduced to expedite the path search process.Then,a traversal pruning strategy is incorporated to reduce redundant paths.Lastly,a directional caching strategy is introduced to utilize the originally planned path as a direction cache,thereby avoiding random sampling.These strategies collectively form the NTDRRT* algorithm,which is an improved version based on the RRT* algorithm.The effectiveness of the NTDRRT* algorithm is verified through benchmark testing.Similarly,to tackle the issues of the ACO algorithm,namely susceptibility to local optima and slow convergence speed,a multi-strategy improved algorithm called HPACO is proposed.Firstly,a heuristic approach to goal strategy is introduced to guide node selection towards the target point,thereby avoiding local optima.Then,a pheromone redundancy evaporation strategy is incorporated to reduce the concentration of pheromones on suboptimal paths,thus improving the convergence speed of the algorithm.These strategies collectively form the HPACO algorithm,which is an improved version based on the ACO algorithm.The effectiveness of the HPACO algorithm is also verified through benchmark testing.Finally,the NTDRRT* algorithm and the HPACO algorithm are combined to form a hybrid improved algorithm,which serves as the basis for an AUV path planning model proposed in this study.Results obtained from three-dimensional simulation experiments demonstrate that the hybrid improved algorithm reduces the number of turning points in AUV path planning by 82% compared to the basic RRT* algorithm,and by 23% compared to the ACO algorithm.Furthermore,it requires 55% fewer iterations to achieve algorithm convergence compared to the ACO algorithm,and 38% fewer iterations compared to the PSO algorithm.Moreover,the planned path length is reduced by 88% compared to the basic RRT*algorithm,by 31% compared to the ACO algorithm,and by 10% compared to the PSO algorithm,which confirms the effectiveness of the hybrid improved algorithm proposed in this study.Additionally,the adaptability of the hybrid improved algorithm is verified through threedimensional experimental simulations conducted in representative terrains.