Navegação robótica em redes de sensores sem fio baseada no RSSI

Abstract

Wireless Sensor Networks (WSNs) are commonly used in monitoring applications due to its capacity to sensing, processing and communicating, and its low cost. However, one limitation of WSN is on energy, because each device (sensor node) of the network needs to have low energy consumption, not allowing the use of extra hardware such as GPS. On the other hand, robots can assist in monitoring made by WSN. One possible application using robots in WSN is to search for events of interest, in which a robot browsing the network to find a specific event, using the signal strength (RSSI) as a reference for navigation. Solutions to this problem have been found in the literature. However, such works assume a devised propagation model, in which the RSSI regression curve versus distance is ideal for that scenario. We present in this dissertation two algorithms that solve the problem of robot navigation based on RSSI in search of an event. The first algorithm is based on the node signal coverage detection and the second uses probability to estimate distance and direction of the target node. Therefore, we conducted experiments to measure the RSSI value according to the distance in Amazon rainforest and represent the signal model propagation obtained in a simulator. Simulations based on the solutions of the literature showed that the percentage of arrival of these solutions is inversely related to the distance of departure from its target when subjected to propagation model detected in experiments. The two algorithms presented have been developed considering the propagation model of the signal obtained in the experiments and both find their target 100 % of the cases.