Sensor deployment by a robot in an unknown orthogonal region: Achieving full coverage

When deploying a wireless sensor network in an unknown environment, commonly referred to as Region of Interest (ROI), the main goal is for the entire region to be covered by the sensing ranges of the deployed sensors. While this goal of full coverage is easily achieved in presence of human intervention, it becomes problematic if the region is dangerous or inaccessible to human. An approach recently proposed to solve the problem is to use a robot to deploy the sensors; the main advantages respect to the alternative of employing mobile sensors are the reduced costs (due to manufacture and maintenance cost of common static sensors vs. mobile ones) and the reduced complexity of the coordination and control algorithms. Indeed several solution algorithms to achieve deployment of sensors by a robot in an unknown region have been proposed in the literature. Unfortunately, even when restricted to orthogonal regions (e.g., city maps, building plans, etc), all the existing algorithms fail to achieve full coverage of the ROI. Specifically, following the existing protocols, the robot would leave uncovered areas near either the boundaries or critical areas (e.g. areas that are linked to the rest of the region by a narrow corridor). In this paper we present an algorithm that overcomes these problems and guarantees that the deployment of the sensors by the robot achieves full coverage in any simply connected orthogonal ROI, whose topology is unknown to the robot. The proposed algorithm has minimal requirements: it does not need GPS but only local orientation by the robot; the communication range of a deployed sensor is limited to its deployed neighbours, and the robot has a similar range; the total number of sensors used is minimal. Also minimal are the robot´s memory requirements, the total amount of robots movements and of communication between robot and sensors.