This paper describes the optimization of hip joint characteristics of a hip disarticulation prosthesis. We attempted to optimize the characteristics for improved ability to walk using our passive walking model, which can walk by utilizing mechanical properties of rigid body segments and joint resistance.
In order to understand how the hip disarticulation prosthesis gait is performed, we interviewed two hip disarticulation prosthesis users. The interviews showed that practical gait in daily life is different from the gait at a training stage. These two types of gaits were named “practical gait” and “training gait.” Users indicated that the training gait velocity was slower than that of the practical gait. Moreover, in the practical gait the heel contact on the prosthesis side was more natural in comparison with the training gait.
Gait measurements showed that the lumbar angle pattern has rapid extension and lateral bending involving the swing prosthesis in training gait. Step length on the sound side is in agreement despite the different types of gait. In practical gait, step length on the sound side agrees with that on the prosthesis side. Gait velocity in practical gait compared with training gait was 28% faster with subject 1 and 7% faster with subject 2. Therefore, practical gait has an improved gait velocity by swinging the prosthesis, as step length on each side is the same. Motion of prosthesis is achieved not by sound lower extremities but by lumbar flexion, extension, and lateral bending. Furthermore, practical gait reduces lumbar motion as much as possible, and reduces muscle force around the lumbar area.
We developed a passive swing model by applying the above characteristics. This model is composed of eight rigid segments: upper torso, pelvis, upper extremities, thigh, shank-foot. Each joint has passive resistance by ligament. The sound hip and lumbar joint have active moments by muscle, which were obtained from measurement. The objective function for practical gait is defined by the following parameters: (1) difference of each step length, (2) amplitude of active moments, (3) difference of cycles between gait patterns and active moments. As these parameters are minimized, postures of segments, translational velocity, angular velocities and cycle, and amplitude of active moment are recorded. In comparison of subjects, calculated motion patterns on the prosthesis side were well in agreement, so this model is available to estimate hip joint characteristics.
When this model simulates a condition of the current hip elastic characteristic weakened by half, the gait velocity is 6% faster and amplitude of lumbar lateral bending moment is reduced 26%. For this reason, weakening current elastic characteristics around the hip joint can easily control the swing of the prosthesis. As a result, adjustment of the hip elastic characteristic can improve the walk capability.
