Vehicle trajectory predicting method based on hybrid dynamic bayesian networks and gaussian process

Abstract

The invention belongs to the technical field of automatic vehicle driving environment cognition and decision-making, and especially relates to a vehicle trajectory predicting method based on hybrid dynamic bayesian networks and gaussian process. According to the method, parameters of MDBN and GP are learned through natural vehicle driving data, and a plurality of vehicle kinematic models are combined through utilizing MDBN, so that short-term trajectory prediction and estimated probabilities of driving intention and driving characteristics are obtained, and then long-term trajectory predictionand representation of uncertainty prediction are conducted through using GP. By adopting the method, short-term prediction characteristics based on a vehicle physical movement model as well as long-term trajectory prediction and representation of uncertainty prediction according to driver information can both taken into account. Compared to an existing vehicle trajectory predicting method, vehicle models, abstract intention and data driving are combined together, and the expansibility of the MDBN model and the GP model are strong, and thus the method is suitable for different driving scenarios and can combine more effective situational information like road information and traffic information.

Description

technical field [0001] The invention belongs to the technical field of environment cognition and decision-making of automatic driving vehicles, and in particular relates to a vehicle trajectory prediction method based on a hybrid dynamic Bayesian network and a Gaussian process. Background technique [0002] At present, a mainstream solution to realize automatic driving of vehicles is based on the "perception-decision-control" architecture. This layered architecture adopts the idea of ​​anthropomorphism, just like people need to perceive the surrounding environment with sensory organs such as eyes, ears, and nose; and then form an understanding and judgment of the surrounding environment through the processing of the perceived information by the brain. So as to make reasonable decisions and plans; finally, through human limbs, such as hands, feet, etc. to perform determined tasks. Obviously, just like the brain is the heart of the human body, the decision-making system of an...

AI Technical Summary

Problems solved by technology

[0004] The first type of method, the vehicle trajectory prediction based on the physical model, only considers the motion characteristics of the current vehicle, which is suitable for trajectory prediction in a short period of time; because the selection of the vehicle physical model directly affects the trajectory prediction results, and the vehicle corresponding to different driving states The physical models are different, so the trajectory prediction cannot be accurately performed according to a single vehicle model; at the same time, because the vehicle driver information (such as driving intention; driving characteristics) is not considered, the trajectory of the vehicle in the long-term range cannot be predicted; in addition, the method The influence of environmental factors around the vehicle on the trajectory of the vehicle is ignored

The second type of method is based on the vehicle trajectory prediction based on driving intention estimation. It can predict the change of the trajectory caused by the vehicle performing a specific operation (such as turning at an intersection: deceleration, steering, acceleration and turning). Some implementation methods can also predict due to Vehicle trajectory changes caused by changes in surrounding environmental factors, so this type of method can predict vehicle trajectory in a long-term range, but because the physical movement characteristics of the vehicle itself are not considered, the prediction error in a short period of time is relatively large

However, due to the need for a large number of calibration data samples, the physical motion characteristics of the vehicle itself are not considered, and the established neural network model has the disadvantages of weak generalization ability and poor interpretability, it is difficult to complete vehicle trajectory prediction in multiple scenarios Find the corresponding reason for the wrong output result, and it is difficult to express the uncertainty of prediction

Images