Decentralized control of three dimensional mobile robotic sensor networks

Abstract

Decentralised control of mobile robotic sensor networks is a fundamental problem in robotics that has attracted intensive research in recent decades. Most of the existing works dealt with two-dimensional spaces. This thesis is concerned with the problem of decentralised self-deployment of mobile robotic sensor networks for coverage, search, and formation building in three-dimensional environments. The first part of the thesis investigates the problem of complete sensing coverage in three-dimensional spaces. We propose a decentralised random algorithm to drive mobile robotic sensors on the vertices of a truncated octahedral grid for complete sensing coverage of a bounded 3D area. Then, we develop a decentralised random algorithm for self-deployment of mobile robotic sensors to form a desired geometric shape on the vertices of the truncated octahedral grid. The second part of this thesis studies the problem of search in 3D spaces. We present a distributed random algorithm for search in bounded three dimensional environments. The proposed algorithm utilises an optimal three dimensional grid for the search task. Third, we study the problem of locating static and mobile targets in a bounded 3D space by a network of mobile robotic sensors. We introduce a novel decentralised bio-inspired random search algorithm for finding static and mobile objects in 3D areas. This algorithm combines the Levy flight random search mechanism with a 3D covering grid. Using this algorithm, the mobile robotic sensors randomly move on the vertices of the covering grid with the length of the movements follow a Levy flight distribution. This thesis studies the problem of 3D formation building in 3D spaces by a network of mobile robotic sensors. Decentralised consensus-based control law for the multi-robot system which results in forming a given geometric configuration from any initial positions in 3D environments is proposed. Then, a decentralised random motion coordination law for the multi-robot system for the case when the mobile robots are unaware of their positions in the configuration in three dimensional environments is presented. The proposed algorithms use some simple consensus rules for motion coordination and building desired geometric patterns. Convergence of the mobile robotic sensors to the given configurations is shown by extensive simulations. Moreover, a mathematically rigorous proof of convergence of the proposed algorithms to the given configurations are given.