Identification of time-varying joint dynamics

System identification is a top-down approach that produces a mathematical model of a system from measurements of its inputs and outputs. These techniques can be used in a wide variety of fields including biomedical systems. Temporal basis expansion methods and ensemble techniques are two such specialized system identification techniques developed for identifying time-varying systems. In this paper, we disclose a novel identification technique for time-varying systems that combines the basis expansion and ensemble approaches. This improves the noise performance vis a vis the basis expansions and reduces the number of realizations needed for accurate identification as compared to the ensemble method. This technique is validated on a Monte Carlo simulation representing the intrinsic compliance of the ankle joint while it moves from full plantarflexion to full dorsiflexion over a period of 2s