A PARAMETERIZED REPRESENTATION FOR SELF-MOTION MANIFOLD OF CRAWLER CRANE ROBOTS

Yuanshan Lin,∗,∗∗,∗∗∗ Fang Wang,∗,∗∗,∗∗∗ Xinzhong Cui,∗,∗∗ Liang Hong,∗,∗∗ and Yanan Liu∗∗∗∗

Keywords

Lift path planning, motion planning, selfmotion manifold, cranerobot, samplingbased planners

Abstract

In lift path planning, the initial/goal is specified with the pose of the lifted object in workspace, while the paths are searched in configuration space (C-space) of crane robots. Due to this mismatch, the initials/goals have to be converted to initial/goal configurations of crane robots before planning. Generally, the initial/goal configurations are determined by inverse kinematics (IK) or Jacobian methods which typically suffer from high computing cost and difficulty in integrating certain constraints. In this paper, we treat the set of initial/goal configurations as the self-motion manifolds of crawler crane robots and present a parameterized representation for it, termed as Crane Location Region (CLR). The CLR representation comprises two parts: the pose of the lifted object and a list of parameters which bound the poses of the crane robot’s base. The initial/goal configurations can be analytically obtained from CLR. Three experiments are constructed to verify the effectiveness of CLR and demonstrate its several features. The experiment results show that the CLR can appropriately represent the self-motion manifolds of crane robots, efficiently generate initial/goal configurations for sampling-based path planners and significantly improve their performance by setting appropriate parameters of CLR. The CLR provides a novel view of generating initial/goal configurations for lift path planners.

Important Links:



Go Back