TrackMe - a low power location tracking system using smart phone sensors

Tracking of locations of a mobile object or person continuously using smart phones using conventional Global Positioning System (GPS) puts a huge toll on the battery life of power-limited smart phones. The power consumption of a GPS unit is much more than any other sensors in a smart phone. Worse, the GPS unit cannot be switched off even if the smart phone is in sleep mode. GPS, in addition, is not effective in indoor locations because suitable number of satellites cannot be obtained for acceptable communication. To overcome these problems, in this work, we have proposed a low power and low cost fast location tracking system for a smart phone device that is effective in tracking continuous locations of a moving device with very good accuracy. The inbuilt sensors of a smart phone like the accelerometer, the magnetometer and the gyroscope have been utilized, instead of the costly GPS unit, to track the continuous locations of a mobile device. Distance covered between any two time instants is calculated using the accelerometer readings, while the magnetometer readings give the direction of movement of the mobile device. These sensors are used to get an initial estimate of the number of foot counts and the average foot length of the person that carries the smart phone. Basically, a pattern has been framed using the change of acceleration due to gravity when a footstep is taken. These values are then used to continuously find the location of a mobile person carrying the device. The system based on these sensor measures has been implemented on Android based smart phone devices. Implementation results have generated an accuracy level of as low as 2 meters distance. The system has been tested on both indoor as well as outdoor locations without any observable differences in performance measures. Huge savings in terms of battery power consumption, as high as 20 percent for a run of 3 hours, have been found, with savings increasing rapidly with the increase in time of run of our system. Our results strongly suggest the use of our proposed system as a good alternative for the costly GPS system of location tracking.