Abstract: A method and apparatus, such as implemented by software code on a mobile device, to estimate a location and a traveling distance by leveraging lower-power inertial sensors embedded in the mobile device as a supplement to the device's GPS. To minimize the negative impact of sensor noises, the invention exploits intermittent strong GPS signals and uses linear regression to build a prediction model which is based on a trace estimated from inertial sensors and the one computed from the GPS. Additionally or alternatively, the invention can utilize landmarks (e.g., bridges, traffic lights, etc.) detected automatically and/or special driving patterns (e.g., turning, uphill, and downhill) from inertial sensory data to improve the localization accuracy when the GPS signal is weak.