Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Coverage First Next Best View for Inspection of Cluttered Pipe Networks Using Mobile Manipulators

Joshua Raymond Bettles, Jiaxu Wu, Bruno Vilhena Adorno, Joaquin Carrasco, Atsushi Yamashita

Abstract

Robotic inspection of radioactive areas enables operators to be removed from hazardous environments; however, planning and operating in confined, cluttered environments remain challenging. These systems must autonomously reconstruct the unknown environment and cover its surfaces, whilst estimating and avoiding collisions with objects in the environment. In this paper, we propose a new planning approach based on next-best-view that enables simultaneous exploration and exploitation of the environment by reformulating the coverage path planning problem in terms of information gain. To handle obstacle avoidance under uncertainty, we extend the vector-field-inequalities framework to explicitly account for stochastic measurements of geometric primitives in the environment via chance constraints in a constrained optimal control law. The stochastic constraints were evaluated experimentally alongside the planner on a mobile manipulator in a confined environment to inspect a pipe network. These experiments demonstrate that the system can autonomously plan and execute inspection and coverage paths to reconstruct and fully cover the simplified pipe network. Moreover, the system successfully estimated geometric primitives online and avoided collisions during motion between viewpoints.

Coverage First Next Best View for Inspection of Cluttered Pipe Networks Using Mobile Manipulators

Abstract

Robotic inspection of radioactive areas enables operators to be removed from hazardous environments; however, planning and operating in confined, cluttered environments remain challenging. These systems must autonomously reconstruct the unknown environment and cover its surfaces, whilst estimating and avoiding collisions with objects in the environment. In this paper, we propose a new planning approach based on next-best-view that enables simultaneous exploration and exploitation of the environment by reformulating the coverage path planning problem in terms of information gain. To handle obstacle avoidance under uncertainty, we extend the vector-field-inequalities framework to explicitly account for stochastic measurements of geometric primitives in the environment via chance constraints in a constrained optimal control law. The stochastic constraints were evaluated experimentally alongside the planner on a mobile manipulator in a confined environment to inspect a pipe network. These experiments demonstrate that the system can autonomously plan and execute inspection and coverage paths to reconstruct and fully cover the simplified pipe network. Moreover, the system successfully estimated geometric primitives online and avoided collisions during motion between viewpoints.
Paper Structure (18 sections, 25 equations, 3 figures, 1 table)

This paper contains 18 sections, 25 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Path Planning for Inspection Tasks
  3. Collision Avoidance in Unknown Environments
  4. Statement of Contributions
  5. Mathematical Background
  6. Whole-Body Kinematic Control
  7. Vector Field Inequalities
  8. Stochastic Vector Field Inequalities for Planes
  9. Coverage First Next Best View
  10. Visual Information Gain
  11. Coverage Information gain
  12. Weighted Information Gain
  13. Sampling and Selection
  14. Experimental Setup
  15. Results and Discussions
  16. ...and 3 more sections

Figures (3)

  • Figure 1: The experimental setup with the constructed mobile manipulator in a single-pipe environment, overlaid with the radiation sensor model represented as a sphere in red, centred on $\mathcal{F}_\mathrm{sensor}$, and the depth sensor model represented by the cameras' FoV in pink.
  • Figure 2: Time response of the closed-loop system from one trial in the three-pipe environment: (a) control input for the base, (b) control input for the arm, (c) norm of the task-space error, (d) distance to planes, and (e) distance to pipe constraints for the mobile manipulator in the confined space. Horizontal red dashed lines represent limits for the associated variables, whereas vertical black dashed lines represent the selection of a new set point.
  • Figure 3: Results of the CFNBV planner within the three-pipe environment with the mean value of the five trials displayed with the maximum and minimum values overlaid. (a) The evolution of the voxels and their categorisation based on the visual and coverage sensors, (b) the minimisation of the combined entropy per voxel, and (c) the interplay between the IGs for environment exploration and surface exploitation.