Home /Research /Rugged - Terrain Traversability Analyzing For Quadruped Robots
LOCOMOTION

Rugged - Terrain Traversability Analyzing For Quadruped Robots

Shangcong Li, Rui Song, Yukun Zheng, He Zhao, Yibin Li

Year
2019
Citations
6

Abstract

This paper presents an algorithm of feasibility analysis of the terrain to improve the adaptability of quadruped bionic robot for arbitrary environments. We use the depth camera to sense the ground environment, and utilize the rasterized Digital Elevation Model (DEM) to process the dense point clouds obtained by the camera to describe the terrain. Through calculating the geometric features in the grid, including slope value, curvature value, roughness and fluctuation, the traversability of terrain could be derived and judged. In this paper, we propose an approach to calculate the roughness characteristics, which increasing the accuracy of calculation and meanwhile adding the calculation of the grid fluctuation. The traversability of the grid is then calculated according to the corresponding formula according to the geometric characteristics of each grid. Finally, a simulation is provided to illustrate the effectiveness of the proposed method.

Keywords

TerrainGridRobotComputer scienceCurvatureAdaptabilityComputer visionPoint (geometry)Process (computing)Artificial intelligence

Related papers

Browse all LOCOMOTION papers