Assessment of WiFi RSS using design of experiment for mobile robot wireless positioning system

Abstract

Wireless infrastructure such as WiFi is a practical option for mobile robot indoor positioning system since many recent building incorporated WiFi solutions in their building. The mobile robot development cost could be greatly reduced especially in the case of multiple robots. The measurement of the WiFi RSS is an essential key point for positioning on fingerprinting approach. Two parameters are concerned which are the WiFi AP antenna heights and the separation distance between the WiFi AP and the receiver. Therefore, it is desirable to see the impact of both parameters in WiFi RSS measurements. In this paper, a series of experiment has been carried accordingly to the design of experiment order. Later, analysis of variance (ANOVA) is performed in order to investigate the most significant parameter for mobile robot positioning system. The result shows that Tx-Rx separation plays the most significant role with zero P-value while the insignificant antenna height parameter is larger than the confident level, α. This observation is later validated by actual mobile robot positioning via fingerprinting technique using the weighted K- nearest neighbor (WKNN) positioning algorithm.