Abstract
This thesis presents a contribution for FES-induced movement in paraplegia. The feasibility of posture estimation from forces exerted voluntarily on the handles of a supporting frame during standing is investigated. The problem is stated as a constraint satisfaction problem and is solved with an algorithm based on interval analysis. The constraints are derived from a kinematic model of the human body. The methodology is capable to take in account any uncertainty in quantities that are either assumed known<br />or measured. Uncertainty bounds are computed for the estimated posture. This method has been first validated with healthy subjects experiments, using two 6-axis force/torque sensors attached onto handles of a supporting frame, and was finally used and validated with paraplegic patients.