Model training method, control method and device of unmanned equipment

Abstract

The invention discloses a model training method and a control method and device of unmanned equipment, and particularly discloses a method for determining a driving scene corresponding to a training sample through a scene driving model for each driving scene and training a decision model corresponding to the driving scene by using the training sample, obtaining each adjusted decision model, determining an actual driving scene corresponding to each training sample on the basis of the adjusted decision model, and training the scene driving model by using the actual driving scene until the scene driving model meets a preset training condition. And then, when the unmanned equipment is controlled, determining a driving scene where the unmanned equipment is located by using the trained scene driving model, then determining a control strategy corresponding to the unmanned equipment by using a decision model matched with the driving scene, and controlling the unmanned equipment to drive. And the adaptive capacity of the unmanned equipment to different driving scenes is improved.

Description

technical field [0001] This specification relates to the field of unmanned driving technology, in particular to a method for model training, a control method and device for unmanned driving equipment. Background technique [0002] Unmanned driving technology is a driving technology in which the vehicle drives itself through various actual scenarios without a driver or the driver does not take over the vehicle. With the continuous advancement of artificial intelligence technology, unmanned driving technology has also made great progress, and various unmanned vehicles are favored by more and more users. At present, there are usually two ways to plan the control strategy in unmanned driving. One is to dynamically plan the driving trajectory based on the data collected by the unmanned vehicle, and then determine the control strategy corresponding to the unmanned vehicle according to the planned driving trajectory. , and then control the driving of the unmanned vehicle through t...

AI Technical Summary

Problems solved by technology

[0003] However, in real life, unmanned vehicles need to face a variety of scenarios, such as mountain roads, off-road roads, urban roads, country roads, etc. In the face of different scenarios, unmanned vehicles need to face The problems and challenges are different, so the focus of attention in the control strategy planning is also different. In the existing control strategy planning methods, most of them are based on the data collected by the unmanned vehicle to determine the corresponding control strategy of the vehicle. , without fully considering the influence of the scene where the unmanned vehicle is located on the decision-making result of the control strategy. Therefore, when the unmanned vehicle is driving in different scenarios, the obtained control strategy cannot be well adapted to the driving environment, and the adaptability of the unmanned vehicle is weak. The problem

Images