Threat Estimation Based Path Planning Method For Intelligent Vehicle

Threat estimation and path planning are the important technology in the fields of intelligent vehicles.Threat estimation to collects and processes environmental information with multi-sensors,and then the environment is apperceived effectively and assesses accurately,and provides a basis for the later decision-making.Path planning is the decision-making process of intelligent vehicles,and on the basis of the perception and modeling of the environment,an effective path is planned.In this paper,the method of threat estimation and the path planning method based on threat estimation is studied.The path planning software is developed and its effectiveness of this method is validated.In order to estimate the target vehicle's threat capabilities,A Bayesian network model is constructed in this paper that includes target characteristics,external environmental factors,and driver factors that affect the vehicle threat index.The environmental factors include road conditions and visibility;the target vehicle conditions include the size of the vehicle,the relative distance between the target vehicle and the host vehicle,the absolute velocity of the target vehicle,and its relative velocity to the host vehicle,the driver conditions include physiological and psychological factors.Then,the network model is built according to the relationship of the influencing factors.In the process of threat estimation reasoning,the Bayesian network model is subjected to variable structure processing to obtain a variable structure Bayesian network model to improve efficiency.Finally,the effectiveness of this paper's threat estimation model is verified by simulation experiments.In this paper,a path planning method based on threat estimation is proposed.Firstly,a multi-valued grid map based on threat estimation is built.According to the target threat index of the occupied grid,different threat fields are formed around the corresponding grid.Each grid's state in the map is determined by the target threat index of the occupied grid or the field of it.Secondly,in the path planning,the grid's state which is lower than the safety threshold can be as a constraint of the path planning to plan an effective path.When there is a lane line as the constraint,the path planning is performed in the sub-map of the local lane.If a feasible path cannot be planned,the adjacent lane sub-map is added to the map and the path is planned to the adjacent lane,thereby reducing the search space to improve the efficiency of path planning.Finally,simulations are used to prove the effectiveness of the proposed method.According to the threat estimation algorithm,path planning algorithm,the intelligent vehicle path planning software is designed and developed,including vehicle threat estimation module,map building module,path planning module,etc.The real vehicle experiment validates the effectiveness of the method and the software system.