Research On Path Planning Based On Ant Colony And Artificial Potential Hybrid Algorithm For Autonomous Vehicle

As an important part of planning&decision system,path planning technology can effectively relieve the pressure of road traffic operation,reduce traffic congestion,and provide satisfactory travel experience for travelers.It can be seen from the research of existing achievements,overwhelming majority of path planning algorithms still have shortcomings.First of all,research of single algorithm for path planning in dynamic and nonocclusive environment is less.Generally,single performance such as the shortest distance or the least travel time is taken as evaluation index,and influence of road attributes,vehicle characteristics and traffic rules are not considered.Secondly,majority of domestic and foreign scholars separate the researching of path planning globally and locally.In fact,they are closely related.Local path planning makes up for the deficiency of global path planning due to ambient noise and poor robustness;global path planning guides local path planning to find the optimal solution.This paper designs a vehicle path planning method based on ant colony and artificial potential hybrid algorithm.Firstly,algorithms and core technologies that related to vehicle path planning are introduced;Analyzing current research status of vehicle path planning domestically and foreign,and clarifing research ideas and contents of this paper.In addition,this paper introduces key technology of achieving environmental information systematically,construction of dynamic road network model and its mathematical description,studies main factors which affect road resistance function deeply in the road network model,and uses the analytic hierarchy process to quantify road resistance function;By optimizing vehicle speed at signal intersection,the increase of traffic time and idle fuel consumption caused by the start and stop of vehicles can be avoided.What's more,vehicle path planning globally is carried out which combined withdynamic road network.The path that planned by improved artificial potential field method is used to guide ant colony algorithm to conduct path searching,which makes up for disadvantage of slow convergence caused by blind searching at the initial stage of algorithm operation;pheromone update strategies globally and locally are optimized and negative feedback channels are constructed to make pheromone update adjust with the change of iteration times adaptively.The hybrid algorithm and dynamic road network are tested and analyzed in sumo.Last but not least,Virtual potential field force model under multiple constraints is established to analyze the impact of different traffic participants on the basis of planned path globally.The improved artificial potential field method is used to optimize local driving path,and effectiveness of algorithm is verified by simulation.Finally,according to proposed ant colony and artificial potential hybrid algorithm and existing experimental conditions,local obstacle avoidance path planning experiments are carried out to verify the accuracy and effectiveness of porposed algorithm.