A Direct Approach of Path Planning Using Environmental Contours
- PDF / 4,569,692 Bytes
- 14 Pages / 595.224 x 790.955 pts Page_size
- 99 Downloads / 182 Views
A Direct Approach of Path Planning Using Environmental Contours Hongjun Yu1
· Lihua Liang1 · Peng Shi2 · Qing Jiang1
Received: 18 February 2020 / Accepted: 5 October 2020 © Springer Nature B.V. 2020
Abstract Roadmap is important in typical robotic applications and it is not a trivial task to obtain in unknown space. In this paper, we propose a novel approach to calculate the roadmap that is robust against noisy environmental contours and the movement of the robot. In order to obtain full visibility to space, we design a direct space partitioning approach to produce the roadmap. It uses readings from rangefinders to establish sequential polygons in time, and as the robot moves, intersections among polygons are iteratively obtained. After iterations of updates, we obtain a number of polygons with stable forms. Based on the connections among the polygons, we obtain a roadmap and propose a routing algorithm to calculate paths between points in space. Simulation examples are provided to demonstrate the performance of the proposed approach. Keywords Space partitioning · Roadmap · Routing algorithm · Environmental contours
1 Introduction Navigation and exploration in unknown space are key issues to address for autonomous robots. Rangefinders are often used for environmental scanning and to obtain raw data. Signals of rangefinders include lidar [18], laser, and ultrasonic. Upon receiving information about the space, autonomous robots initiate mapping and positioning. These actions are critical in calculating roadmaps and planning paths. The results are used by autonomous robots to complete various tasks, such as delivery, assembly and unmanned exploration. Researchers have focused on how to efficiently interpret and utilise the space for autonomous robots. Various research focus on utilising key points in space for path planning. Obstacle-surrounding vertices on artificial grids are used in [1] to generate collision-free path for mobile robot. An improvement in [2] uses ratings of obstacles to reduce computational cost. Waypoints are used in [3] to reduce exposure to radioactive sources, and the results demonstrate advantage over A* algorithm in computation time. Multi-objective optimization under spatial constraints Hongjun Yu
hongjun [email protected] 1
College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin, 150001, China
2
School of Electrical and Electronic Engineering, University of Adelaide, Adelaide, 5005, Australia
in [6] are used for path planning to strike a tradeoff between path complexity and computational efficiency. The methods mentioned above are mostly based on grid or image pixels. This introduces scaling issues and the fixed spacing could lead to uneven performance in unevenly distributed space. Our approach does not impose these restrictions, avoiding the issues while saving computation load. Methods based on Markov decision process often uses sequential data to obtain smooth paths. Sampling-based online re-planning is used in [4] to utilise only local infor
Data Loading...
