Representation of 3D Structure for Path Planning with Biped Wall-Climbing Robots

Abstract

For biped wall-climbing robots, mapping and extraction of the parametric expression map denote challenging tasks, especially for the environments where lidar measurements degrade. To build a representation of the environment for global path planning, a three-dimensional (3D) planar structure modelling and parameterization system is proposed. The system includes the two following tasks: 1) mapping by a lidar-IMU fusion method in which point cloud registration algorithm tracks the lidar poses with the initial transformation matrix provided by IMU integration; 2) extracting the vertexes of planes that represent the parametric expression of the 3D planar structure. Experiments showed that our mapping algorithm was more rubust than others in the lidar-degraded environment. Besides, the precision analysis verified that the parametric expression map was accurate enough to be used for path planning of biped wall-climbing robots.