A K-NN based Area Positioning System in Wireless Sensor Networks

In a wireless positioning system, the signal transmission process is susceptible to environmental influences, causing signal distortion or attenuation, which will affect the accuracy of positioning. To reduce the positioning error caused by the poor signal, most of the environmental models are established using pattern matching, which can effectively reduce the positioning error caused by the unstable signal. This paper utilizes a K-nearest neighbor (K-NN) algorithm to enhance indoor area positioning accuracy for Wireless Sensor Networks (WSNs). Simulation results show that the key factor influencing positioning accuracy is the number of positioning reference points used in the algorithm. When the number of positioning reference points increases from 4 to 5, the highest average correct rate in the simple environment rises by 9.24% from 75.2% to 84.44%. Similarly, in the complex environment, it increases by 5.80% from 73.64% to 79.44%.