Recently, a number of promising path planning algorithms, including sampling-based RRT, grid-based Hybrid-A*, have been proposed to navigate autonomous vehicles in complex environments. However, the drawbacks of these algorithms make them insufficient for path planning when used separately in real-world traffic scenarios. Thus, it is crucial to build a unified framework that can combine the merits of different path planning algorithms. In this paper, a general and practical path planning framework for autonomous vehicles is proposed. At first, a sampling-based algorithm is designed to plan simple driving maneuvers such as following or changing traffic lane, or overtaking frontal obstacles. When complex maneuvers are required for free space navigation, which is beyond the scope of sampling-based algorithms, a grid-based algorithm is proposed to navigate the vehicle toward the target position along a collision-free trajectory. Off-line simulation results illustrate the generality and practicability of the proposed framework, which can effectively inherit the advantages from both sampling-based and grid-based path planning algorithms, and circumvent their individual shortages in the meantime. The proposed path planning framework was applied to our intelligent vehicle platform ‘Challenger’, which took part in the 7th ‘Intelligent Vehicle Future Challenge’ autonomous vehicle competition in China.
A general and practical path planning framework for autonomous vehicles
2016-11-01
685582 byte
Conference paper
Electronic Resource
English
PATH PLANNING FOR AUTONOMOUS AND SEMI-AUTONOMOUS VEHICLES
European Patent Office | 2021
|PATH PLANNING FOR AUTONOMOUS AND SEMI-AUTONOMOUS VEHICLES
European Patent Office | 2021
|PATH PLANNING FOR AUTONOMOUS AND SEMI-AUTONOMOUS VEHICLES
European Patent Office | 2021
|PATH PLANNING FOR AUTONOMOUS AND SEMI-AUTONOMOUS VEHICLES
European Patent Office | 2023
|Path planning for autonomous and semi-autonomous vehicles
European Patent Office | 2021
|