Unmanned vehicle track real-time planning method based on road structure weight fusion

Abstract

The invention provides an unmanned vehicle track real-time planning method based on road structure weight fusion, which is suitable for intelligent track planning of an unmanned vehicle in an autonomous driving process. The technical scheme adopted by the invention comprises the following steps: 1. simplifying the calculation of a vehicle motion equation by a table lookup mode; 2. adopting a structured and unstructured environment weighting method based on breadth-first search; 3. conducting a dynamic target area decision; 4. adopting an improved A* path search algorithm fused with vehicle dynamics; and 5. conducting an incremental track planning of a dynamic scene. The unmanned vehicle track planning problems under various environments are unified by using a structured and unstructured environment weighting method; the continuity and the stability of the track are realized by utilizing the incremental track planning of a dynamic scene, and the running stability of the unmanned vehicleat low, medium and high speeds is ensured; and a vehicle motion equation and a collision detection process are simplified and approximated, the efficiency of a track planning algorithm is effectivelyenhanced by avoiding floating point operation, the track planning algorithm is enabled to achieve nearing real time, and the correctness of track planning is guaranteed.

Description

technical field [0001] This application is a general real-time trajectory planning method that can be applied to unmanned vehicles in structured and unstructured driving scenarios. It is suitable for intelligent trajectory planning of unmanned vehicles in the autonomous driving process of expressways, urban roads, parking lots and other areas. Background technique [0002] The trajectory is the result of the intelligent decision-making planning system of the unmanned driving system. The unmanned vehicle performs tracking control according to the planned trajectory, so as to realize autonomous driving. Most of the current unmanned vehicle decision-making planning systems need to distinguish between structured and unstructured driving scenarios. In the former scenario, multiple candidate trajectories are generated based on map-based predetermined paths and vehicle dynamic constraints, and then real-time choose. This method has high efficiency and low technical difficulty, b...

AI Technical Summary

Problems solved by technology

This method has high efficiency and low technical difficulty, but it needs to rely on accurate lane information, and it is difficult to apply in non-institutionalized driving scenes such as squares and other areas

Moreover, this method is difficult to achieve optimal driving decisions in complex driving scenarios due to its greedy strategy.

In unstructured scenarios, existing unmanned driving systems generally use search algorithms such as A* and RRT for planning on grid maps, but their real-time performance is generally poor, and it takes seconds to generate autonomous driving trajectories , so it is only suitable for low-speed scenarios, and cannot be applied to medium and high-speed driving on normal structured roads

In addition, the trajectory planning method that distinguishes between structured and unstructured scenarios needs to switch planning strategies online based on high-precision maps or perception information, which reduces the intelligence of the decision-making process for unmanned vehicles.

In summary, there is currently no real-time trajectory planning method that can meet the needs of autonomous driving of unmanned vehicles in structured roads and unstructured environments.

Images