Download files
Abstract
The self deployment of a decentralized mobile sensor network has been addressed in this thesis. The mobile robotic sensors are self deployed in the form of a blanket over a bounded and connected unknown two dimensional region. The main objective is to fully track and to cover the arbitrary unknown region (meeting assumptions) with minimum number of mobile robotic sensors moving under decentralized control. The issue of connectivity has also been considered during final deployment of mobile robotic sensors. An algorithm with decentralized control law has been developed to self deploy the mobile robotic sensors into a triangular lattice pattern ensuring optimization. The optimization of deployment in terms of minimum number of sensors ensuring 1-coverage and 6-connectivity (wherever possible) has been considered during final deployment of mobile robotic sensors over the unknown region. A special consideration in continuation to the past work has been taken in order to remove the redundant placement of mobile robotic sensors on the boundary of the region and to fully track all the segments of the boundary of the region. The control action for velocity and heading of mobile robotic sensors is based on simple nearest neighbourhood rules derived from swarm intelligence. The algorithm also uses the concept of stable quotient system of the decentralized control system during line forming mode inside the region. A number of numerical simulations while changing the arbitrary region have been done in order to validate the control algorithm. As compared to the past work, the boundary moving mobile robotic sensors fully explore every segment of the boundary of the unknown region, and redundantly placed mobile robotic sensor on each of the line forming mode inside the region has been removed.