Placement and Relocation of Wireless Sensor Nodes by a Team of Robots

Abstract

We propose a novel localized sensor placement algorithm for wireless sensor networks, named Back-Tracking Deployment (BTD). Through simulation we show that BTD far outperforms the only competing algorithm LRV [2] in robot moves and robot messages at no additional cost in a failure-free environment and at minimal cost of sensor messages in a failure-prone environment. Further, we propose the first localized carrier-based sensor relocation algorithm for wireless sensor networks, named Randomized Robot-assisted Relocation of Static Sensors (R3S2). We also presented a grid-based variant, called G-R3S2 and a cluster-based variant, called C-R3S2 and C-G-R3S2. Through extensive simulation, the effectiveness and practicality is shown. The simulation results indicate in particular that robots following the LRV policy (G-R3S2 and C-G-R3S2) outperformed those that did not across all measured metrics. Further, the cluster-based variants outperformed their counterpart (C-G-R3S2 vs. G-R3S2 and C-R3S2 vs. R3S2) across all metrics.