Research On The Kinematic Model Of Low Limb Joint For The Control Of Prosthesis

In order to return to the society, amputees need to wear prosthesis to replace the function of intact limb. Compared with traditional prosthesis, which was simple, heavy and inconvenience, active intelligent prosthesis has shown great advantages. Not only can they enable the amputees to walk again, but also improve their life quality. With the rapid development of science and technology, new manufacturing method and materials continue to emerge, making the technology of traditional prosthesis a crossed technology which combined with subjects like material, mechanical, electronic, computer, biological and medical. The structure and control method of artificial limbs are two important parts of key technology of prosthesis. This paper studied on the prediction of motion model which is important to the control of prosthesis. The aim of this study is to find a method to reconstruct a knee joint motion model, which can adapt to different people and different walking-speed. This paper started the research as following:(1) The collection of motion data of lower limb: collecting the lower limb motion data of normal volunteers walking in different speeds on level floor, which contains the angle data of three joints(hip, knee and ankle) and the reflective force.(2) The construction of knee joint motion model: the knee angle curve in a stepping period was analyzed in two methods(phase polynomial-fitting and overall Fourier series fitting). The relationship between walking speed and parameters in the fitting function was studied and the speed adapting knee joint motion model was constructed, which would be used to control the prosthesis.(3) The experimental verification of knee joint motion model based on Fourier series: On the experimental platform of knee joint prosthesis prototype, an experiment was designed to verify the Fourier series model. The force sensor data of intact limb’s plantar was collected to get the walking speed and the motion curve of the prosthesis was calculated. At the same time, the motion data of knee angle was get from the wearable angle sensor. The performance of the motion model was verified by combing the prototype motion curve and measured curve.Through the study, we found that in the Fourier series model, which fits the motion curve of knee angle in a step period, the fundamental angular frequency ω is proportional to stepping period and the ample of rest sinusoidal and cosine component differs according to different persons, but has the same trends. Therefore, we proposed a method to reconstruct knee trajectory which is called reference-offsets. An offset value was added on the base of someone’s knee joint motion model to get other knee joint motion model. The simulation analysis and test on actual prosthesis both showed that the knee joint trajectory reconstructed by the reference-offsets can adapt to the shifty walking-speed and the method can adjust the offsets to meet the different people according to the characteristics of different people.