Human gait analysis using instrumented shoes

This paper describes a human gait analysis and characterization system, that automatically assesses the severity of gait disorders, using low cost 3D instrumented shoes. Human gait pattern analysis has an increasing importance in rehabilitation medicine, sports and other areas. Therefore, understanding gait patterns enables medical staff to follow recovery processes of patients and adjust their treatments, as well as it permits to evaluate the efficiency of athletes' gait in order to improve their performances. At the present time, there are many systems that measure some parameters that characterize human gait, like the vertical component of the ground reaction force (GRF), which is used to map the plantar pressure and to compute the centre of pressure's location (CoP). Most present solutions use force plates and insoles with pressure or force sensors [1-3]. However, force plates restrict the number of steps that can be done and sometimes insoles don't adjust very well to the footwear. The presented computerized system consists of a pair of low cost instrumented shoes designed to read and collect the three components of the GRF during the human gait using thin force sensors. Data is sent to a computer via a wireless protocol. They can be processed and visualized in graphs showing the three curves of the GRF and the trajectory of the CoP. These instrumented shoes have already been tested in persons with and without physical disorders to build a gait database. These persons had different ages, weights and heights, and walked at five different velocities (slow, very slow, normal, fast and very fast). Time curves of the GRF, biometric parameters and the walking velocity are used as inputs to a neuronal network in order to generate reference gait patterns for people with different physical characteristics. Results were consistent with those in the literature. Some tests have been done in a public hospital, with patients subjected to ligamentoplasty two years ago, because of the rupture of their knee's anterior cruciate ligament. The aim of the current work is to infer their recovery degree, comparing their GRF and CoP curves with the reference gait patterns previously obtained. Preliminary results show that it is possible to quantify differences in gait patterns of non-healthy people. The presented system can thus be an important gait disorder diagnostic tool as it objectively quantify gait disorders, something that is harder to get with the present subjective analysis.