RegTrack: a differential relative gps tracking solution.

ABSTRACTIn many mobile wireless applications such as the automated driving of cars, formation flying of unmanned air vehicles, and source localization or target tracking with wireless sensor networks, it is more important to know the precise relative locations of nodes than their absolute coordinates. GPS, the most ubiquitous localization system available, generally provides only absolute coordinates. Furthermore, low-cost receivers can exhibit tens of meters of error or worse in challenging RF environments. This video demonstration presents an approach that uses GPS to derive relative location information for multiple receivers. Cooperating nodes in a network share their raw satellite measurements and use this data to track the relative motions of neighboring nodes as opposed to computing their own absolute coordinates. This is achieved via creation of a new error model that incorporates GPS localization errors specific to the multiple-receiver case, development of a new, highly accurate observation model that allows for the change in range between a single satellite and two receivers to be mapped through time, and the synthesis of these two models into a novel 3D pairwise tracking algorithm that uses the local GPS node as its own reference and requires only an approximate estimate of its own absolute position to achieve centimeter-scale relative tracking precision. The system just described was implemented on a network of Android smartphones equipped with a Bluetooth-enabled, custom GPS node to provide raw measurement data. Several experiments were carried out to test our proof of concept in various GPS environments and under different types of dynamic conditions. Our evaluation shows that centimeter-scale tracking accuracy at an update rate of 1 Hz is possible under various conditions with the presented technique. This is more than an order of magnitude more accurate than simply taking the difference of reported absolute node coordinates or other simplistic approaches due to uncorrelated measurement errors. The demo shown in this video represents one of the high-dynamic test cases in which we placed a stationary GPS receiver on a tripod in the middle of a parking lot and three additional nodes on top of an automobile. We proceeded to drive the automobile through various environments, including a multipath-rich alleyway, moderately benign suburban roads, and at high speeds on an interstate highway. The total length of the experiment was 15 minutes, spanning 12.2 km of terrain, with the baseline vector between the stationary node in the parking lot and the roving nodes on the automobile ranging from 0 to 3.5 km at any given time. The visualization of the experimental results is taken from the point of view of the stationary node, enabling us to map the relative tracks of the roving nodes in an absolute coordinate space using Google Earth. The results of this experiment showed centimeter-scale tracking accuracy - a level of precision that enables clear distinction of driving features such as lane shifts and velocity changes - under various driving conditions, with only one problematic loss of satellite locks due to signal obstruction by a wide overpass while the vehicle was changing directions. This momentary loss of lock caused the tracking algorithm to resume after approximately 2 meters of error had already been accrued, and this error remained present and constant throughout the remainder of the experiment. Future work involves solving for initialization and re-calibration of the tracking algorithm on the fly, such that losses of lock and visibility obstructions do not degrade the accuracy of the system in the long run. In its present form, RegTrack already provides an order of magnitude better precision than complementary methods using low-cost, single-frequency commercial receivers, and the stage is set for further research to increase its robustness and utility for high-precision, low-cost applications.