A control method and device for an unmanned driving device

Abstract

This specification discloses a control method and device for unmanned driving equipment. Involved in the field of unmanned driving technology. The target obstacle is determined according to the intersection track of the obstacle and the unmanned equipment. According to each track point of the unmanned driving device and its current speed, determine the first vector, and for each target obstacle, determine the second vector according to the position of the target obstacle within the preset time and the speed of each position. vector. Input the two vectors into the pre-trained decision model to get the decision corresponding to the target obstacle. A control strategy for the unmanned device at the next moment is determined according to at least part of the obtained decisions, so as to realize that the unmanned device and each target obstacle pass through corresponding trajectory intersection positions. It can predict the intersection of trajectories and determine the control strategy of the unmanned equipment at the next moment based on the decision, which can avoid the simultaneous occurrence of unmanned equipment and obstacles at the intersection position, avoid collisions, and ensure the safety of unmanned equipment.

Description

technical field [0001] This specification relates to the field of unmanned driving technology, and in particular to a control method and device for unmanned driving equipment. Background technique [0002] Usually, when the unmanned driving equipment is driving automatically, the control module in the unmanned driving system controls the driving of the unmanned equipment to avoid obstacles in the environment according to the next decision of the unmanned equipment determined by the decision-making module Objects, especially obstacles that need to be avoided and which may have a risk of collision with the unmanned equipment, so that the unmanned equipment can drive safely. [0003] In the existing technology, when determining the next decision of the unmanned equipment, the state decision set determined based on the driver's driving data is pre-constructed, that is, the decision made by the unmanned equipment and its environment and the driver Formed collection, and store th...

AI Technical Summary

Problems solved by technology

The design of the reward function is more difficult. If the reward function is unreasonable, it may lead to inaccurate decision-making of the unmanned equipment based on the deep reinforcement learning network, or even a major mistake, causing the unmanned equipment to be in danger.

Images