A Learning-based Approach to the Real-time Estimation of the Feet Ground Reaction Forces and Centres of Pressure in Humans

The feet centres of pressure (CoP) and ground reaction forces (GRF) constitute essential information in the analysis of human motion. Such variables are representative of the human dynamic behaviours, in particular when interactions with the external world are in place. Accordingly, in this paper we propose a novel approach for the real-time estimation of the human feet CoP and GRFs, using the whole-body CoP and the human body configuration. The method combines a simplified geometrical model of the whole-body CoP and a learning technique. Firstly, a statically equivalent serial chain (SESC) model which enables the whole-body CoP estimation is identified. Then, the estimated whole-body CoP and the simplified body pose information are used for the training and validation of the learning technique. The proposed feet CoP model is first validated experimentally in five subjects. Then, its real-time efficacy is assessed using dynamic data streamed on-line for one selected subject.