A Particle Filter Approach for Pedestrian Dead Reckoning Using Wearable Sensors

In this paper, a pedestrian dead reckoning (PDR) method is proposed based on the particle filter framework with map information measurements from both an inertial measurement unit (IMU) on the footwear and a global positioning system (GPS). The main challenging of computing PDR with IMU signals is that the biases of IMU sensors will result in unbounded integration drift. Therefore, a new method that uses non-zero velocity data for zero-velocity update of the accelerometer biases is proposed. Also, a measurement equation for the rate-gyro biases using accelerometer measurements are derived. We also show that some heuristic drift canceling rules can be easily incorporated into the particle filter framework. Experimental results show that the proposed method can effectively reduce the accumulation error from 42 meters to 10 meters.