An intelligent multi-sensor system for first responder indoor navigation

Abstract

This paper presents an indoor navigation system based on sensor data from first responder wearable modules. The system combines an inertial measurement unit, a digital camera and a radio frequency identification device in a way that allows the advantages of each sensor to be fully exploited. The key to this synergy is the extracted qualitative criteria which characterize the performance of each sensor subsystem at various first responder activities and operational conditions under certain time intervals. The accuracy of the detected walking pattern through measurements of the acceleration magnitude from the inertial sensor is utilized for the performance evaluation of the dead-reckoning algorithm. The amount of correct feature matches is linked to the three-dimensional scene representation from the camera navigation subsystem and finally, the degree of probability of each radio frequency identification location estimate is exploited as a straightforward qualitative criterion. The final fused location estimation is extracted after applying fuzzy if–then rules at each time interval. Since the inertial sensor suffers from accumulated drift, the rules of the fuzzy inference system drop the measurements from the inertial measurement unit whenever the other two subsystems perform adequately. Extensive comparison and experimental results based on the proposed architecture have shown not only better navigation effectiveness and lower positioning error compared with other first responder navigation systems but also increased accuracy in various and challenging operational conditions.