Traditional Received Signal Strength Research Articles

RSS-Based Indoor Localization Algorithm for Wireless Sensor Network Using Generalized Regression Neural Network

Traditional received signal strength (RSS)-based localizations are often erroneous for the low-cost WSN devices. The reason is that the wireless channel is vulnerable to so many factors that deriving the appropriate propagation loss model for the WSN device is difficult. We propose a flexible location estimation algorithm using generalized regression neural network (GRNN) and weighted centroid localization. In the first phase of the proposed scheme, two GRNNs are trained separately for x and y coordinates, using RSS data gathered at the access points from the reference nodes. The networks are then used to estimate the approximate location of the target node and its close neighbors. In the second phase, the target node position is determined by calculating the weighted centroid of the Nc-closer neighbors. Performance of the proposed algorithm is compared with some existing RSS based techniques. Simulation and experimental results indicate that the location accuracy is satisfactory. The system performance is remarkably good in comparison with its simplicity and requiring no additional hardware.

A Newton Iteration Based WSN Localization System in Underground Tunnels

One of the main concerns in under construction tunnels is ensuring the safety of the workers and their equipment. Being aware of the real-time position of personnel in such harsh environment is challenging and usually requires a sophisticated localization system. With traditional Received Signal Strength Indicator (RSSII) failing to accurately estimate the distance between nodes due to multipath effect in such long and narrow space, Radio Frequency Time-of-Flight (RF-TOF) is proved to be an alternative method for more accurate distance estimation. To reduce the communication cost, a distributed localization system is proposed, where a simple Newton Iteration location estimation algorithm is embedded in the blind node. Linear least square estimation is used as the initial value to accelerate the convergence of the iteration. Experimental results show the effectiveness of the proposed system