Gait analysis is the analysis of human movement used to assess the way we walk or run, and requires accurate measurements to model biomechanical aspects of the actual movement. Gait analysis provides information about the nature of movement disorder to neurologists and rehabilitation experts, which is useful for deciding appropriate medical intervention. In gait analysis, one computes the translation and rotational position of limbs from the measurements. A recent trend has been the use of Inertial Measurement Units (IMU) to measure gait. However, IMU based angle and position computations are plagued by sensor noise, and result in the well-known position (angle) error called drift (a continuously increasing deviation away from the true position). There are several known techniques for reducing drift. An advanced technique is the use of a Kalmanfilter, which estimates the sensor noise and attempts to remove that from the position computations. A much simpler, and accurate, method exists called the ZUPT - or the Zero Velocity Update algorithm. The algorithm checks for stationary (zero velocity) conditions of a gait and applies correction to the position computations. This error mitigation helps in accurate position determination of a body. ZUPT is a well-known technique and this paper presents a modification to the classical ZUPT algorithm along with detailed real time experimental results of the same. We combine ZUPT with a high pass filter to obtain higher accuracies. Future work will incorporate the use of body kinematics to determine when the body will be at rest (during a gait cycle) and update the velocity using the ZUPT.