Target obstacle vehicle trajectory prediction method based on Bayesian network

Abstract

The invention discloses a target obstacle vehicle trajectory prediction method based on a Bayesian network and belongs to the field of driverless vehicles. An environment model calculates a historicalpath probability PA, a course angle probability PB, a target obstacle vehicle and other obstacle distance influence probability PC, a historical speed probability PD, a historical obstacle avoidancespeed probability PE and a target obstacle vehicle speed probability PF and inputs the probabilities into a Bayesian network model; with the probabilities PA, PB, PC, PD, PE and PF taken as the childnode input of a Bayesian network model, and a next path grid probability and a next grid speed probability as root nodes, a Bayesian network structure is constructed; different posterior probabilitiesfrom the current grid to surrounding eight grids and the posterior probability of the speed corresponding to each grid are obtained; and a grid and a speed which have the highest posterior probability are obtained through comparison, and are adopted as the next trajectory point of a target obstacle vehicle. According to the method, the influence of the distance between the target obstacle vehicleand other obstacles around the target obstacle vehicle in a dynamic environment is considered, and the predicted future trajectory is more accurate.

Description

technical field [0001] The invention relates to the field of driverless cars, in particular to a method for predicting the trajectory of a target obstacle vehicle driving around the driverless car. Background technique [0002] In the field of unmanned driving, it is necessary to plan the trajectory of unmanned vehicles to obtain a safe and smooth running trajectory. In order to enable unmanned vehicles to avoid surrounding obstacles more accurately, it is usually necessary to predict the future trajectory of the obstacles. Generally speaking, the trajectory points of the obstacle at multiple preset times in the future are predicted according to the movement state of the obstacle in the current and previous preset time periods, so as to obtain the future trajectory of the obstacle. However, the accuracy of the future trajectory of obstacles predicted in this way is poor, which makes the results of unmanned vehicle trajectory planning lack accuracy. For example, in the docu...

AI Technical Summary

Problems solved by technology

However, the future trajectory accuracy of obstacles predicted in this way is poor, making the results of unmanned vehicle trajectory planning lack accuracy

For example, in the document with the Chinese patent application number 201910034446.9, an unmanned driving trajectory planning method based on obstacle prediction and MPC algorithm is published, which obtains the movement speed and direction of surrounding vehicles through on-board sensor equipment and corresponding image recognition algorithms. It assumes that in the prediction time domain, the surrounding vehicles are traveling along the current lane at the current speed, and the surrounding vehicle trajectory is obtained, which ensures the safety of its later trajectory planning. The impact of other surrounding obstacles on its driving is not consistent with the actual dynamic driving environment

Images