This paper proposes an affordable mobile platform for pathological gait analysis. Gait spatio-temporal parameters are of great importance in clinical evaluation but often require expensive equipment and are limited to a small and controlled environment. The proposed system uses state-of-the art robotic tools, in contrast to their original use, for the development of a robust low-cost diagnostic decision-making tool. The mobile system, which is driven by a Kinect sensor, is able to (1) follow a patient at a constant distance on his own defined path, and (2) to estimate the gait spatio-temporal parameters. The Robust Tracking-Learning-Detection algorithm estimates the positions of the targets attached to the trunk and heels of the patient. Real-condition experimental validation including the corridor, occlusion cases, and illumination changes was performed. A gold standard stereophotogrammetric system was also used and showed good tracking of the patient and an accuracy in the stride length estimate of 2%. Finally, preliminary results showed an RMS error that was below 10°in the 3D lower-limb joint angle estimates during walking on a treadmill.