The evolution of smartphones and their embedded sensors motivates research toward the development of handheld device based navigation solutions especially for harsh environments. In this context, Pedestrian Dead Reckoning is usually adopted to compute the pedestrian's trajectory. Step/stride lengths and walking directions are combined in a recursive process. Unfortunately the estimated path suffers from drifting errors due to the sensors' nature and the motion complexity. To reduce this error, map matching strategies are studied and several solutions are proposed in the literature. In this work a Matching Filter is proposed to mitigate the drifting errors. The Matching Filter is a nest filter based on an extended Kalman Filter and a Complementary filter. The key idea is to match the PDR trajectory with the standalone GPS trajectory during opportune phases in order to estimate a global heading and scale factor errors on the PDR path. The proposed strategy is tested with a 1km walk in a shopping center. A 75% improvement is found as compared to the PDR only trajectory.