Estimating Walking Distance with a Smart Phone

The global positioning system (GPS) is widely used for localization. However, GPS does not perform well in an indoor environment due to that it is hard to receive satellite signal inside a building. A huge body of work utilized signal strength of short range signal (such as WiFi, Bluetooth, ultra sound or Infrared) to build a radio map for indoor localization, by deploying a great number of beacon nodes in the building. The drawback of such an infrastructure-based approach is that the deployment and calibration of the system is costly and labor-intensive. To overcome that, some prior studies proposed the use of Pedestrian Dead Reckoning (PDR) for indoor localization. The PDR system does not require to build a beacon-based infrastructure, in which a small number of sensors are put on the pedestrian. These sensors (such as G-sensor and Gyro) are used to estimate the distance and direction that the user traveled. The PDR approach can be generally categorized into two types: foot-mounted and waist-mounted. In general, the foot-mounted system can get accurate step length, but perform poorly in estimated heading direction. On the other hand, the waist-mounted system can estimate direction with high accuracy, but is hard to measure the step length. In this work, we proposed a waist-mounted based PDR using one 3-axis accelerometer and one gyroscope sensor. We utilize vertical acceleration to implement double integral for measuring the userâs instant height change and use some physical features of vertical acceleration during the walking to calibrate the measurement. Then based on the Pythagoras´ Theorem, we can estimate each step length based on the userâs height change during his/her walking. Our experiment results show that the accuracy of placing smart-phone on the waist is about 97.35% and placing smart-phone on the chest pocket is about 96.14% in estimating the userâs walking distance.