Collaborative localization in Wireless Sensor Networks based on dependable RSSI

Localization in WSNs (Wireless Sensor Networks) is an active research topic, and many localization algorithms in WSNs have been proposed in the literature. Among them, approaches based on RSS (Received Signal Strength) are popular and useful for distance estimation, since it is simple and needs no extra hardware. However, RSS is greatly affected by the environment and cannot be precisely measured. To tackle this problem, this study proposes a mechanism for distance estimation using dependable RSSI (Received Signal Strength Indicator) values. The threshold for selecting dependable RSSI values is determined by practical experiments. In the proposed approach, each node periodically broadcasts packets to its one-hop neighboring nodes and the neighboring nodes measure the RSSI values of the received packets. The dependable RSSI value is then used to estimate the distance between two nodes. By using the shortest path algorithm, distances between the blind node and the reference nodes can be derived with high accuracy. Thus, for a blind node, with distance information to at least three reference nodes, its location can be computed by the Minimum Mean Square Error algorithm. The proposed localization scheme is implemented with TI CC2430/2431 chips. Practical experiment results show that, with the proposed scheme, the average location error of the blind node is 0.76 meter in an indoor environment under the condition of the dependable RSSI threshold being -59.26 dBm, and 5.91 meters in an outdoor environment under the condition of the dependable RSSI threshold being -79.77 dBm.