LAUSR

Recent research showed that human mobility is characterized by reproducible patterns, i.e., humans tend to travel a few known places. Timely identification of these significant journeys has prospects for emerging intelligent applications like real-time traffic route recommendation and automated HVAC systems. Existing mobile systems, however, utilize energy-hungry sensors like GPS and gyroscope to detect significant journeys, which make it hard to keep such systems running to continuously monitor driving routes. To address this issue of energy efficiency without compromising the performance, in this paper, a hybrid mobile system based on the barometer sensor of a smartphone is developed. Distinctive elevation signatures of driving routes are captured using the smartphone barometer sensor that is exceptionally energy-efficient and position/orientation-independent. Degraded accuracy due to flat areas with minimal elevation changes is offset by developing an adaptive algorithm that opportunistically obtains GPS locations for a very short period of time when such flat areas are detected in real time. Using over 150 miles of field data, it is demonstrated that the proposed mobile system achieves the mean detection accuracy of 97% with the mean false positive rates of 1.5%.