Open AccessAn improved dynamic window approach for local trajectory planning in the environment with dense objects
Author(s) -
Xiquan Mai,
Di Li,
Jian Ouyang,
Yongchao Luo
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1884/1/012003
Subject(s) - obstacle , computer science , motion planning , trajectory , heading (navigation) , window (computing) , robot , path (computing) , acceleration , field (mathematics) , obstacle avoidance , artificial intelligence , computer vision , simulation , mobile robot , engineering , mathematics , aerospace engineering , geography , physics , archaeology , classical mechanics , astronomy , pure mathematics , programming language , operating system
The Dynamic Window Approach (DWA) has been one of the most popular solutions in local trajectory planning due to the advantages of movement fluency. However, the traditional DWA faces the challenge of low-efficiency in the case of local trajectory planning since the robot cannot perceive the density of the obstacle. In this paper, we propose an improved DWA to solve this problem. First, we use multi-sensor technology and new evaluation algorithms to enable the capability of density perception for the disorderly environment. Second, the evaluation factor related to the density change is introduced into the path evaluation function, which will facilitate the robot to perceive the distribution of dense objects in advance. Computer simulation and field experiments show that the efficiency of the improved DWA is increased by 25%. The improved DWA not only considers the original factors such as heading angle, but also introduces a density factor to evaluate the next path, so as to avoid entering dense areas in advance. It can be seen that the improved DWA has relatively stable speed and acceleration, and can avoid dense areas in advance, which can be widely used in robots running in dense environments.