The interfacial pressure between an amputee's residual limb and prosthetic socket is an important clinical issue in the prosthetic fitting process due to the problems associated with poorly fitted sockets. There are two widely studied methods currently available for measuring and monitoring the limb/socket interfacial pressures, experimentally using transducers and numerically using Finite Element Analysis (FEA) which have both been recognised as having limitations. Therefore, a more practical, less invasive and passive sensing approach is needed for this type of investigation if such research is to lead to the production of a practical tool to aid prosthetists in their assessment. An Artificial Neural Network (ANN) and experimental/numerical data has been investigated and combined to develop a Hybrid Inverse Problem Engine (HIPE) for the prediction of the interfacial pressure between a prosthetic socket and residual limb using sensors connected to the socket external surface. A diagnostic socket was manufactured and fitted to a volunteer subject from a computer tracing of the limb. The socket was divided into sixteen patches (or regions) and fifteen strain gauge rosettes were attached. The limb geometry tracing was also utilised to generate a FEA model of the socket and was divided into identical patches which allowed a study to find the key regions of the socket which were sensitive to pressure applied within it. This was performed to minimise the size of the ANN (i.e. reduce the number of required inputs). The final HIPE was found to be able to predict the position and magnitude of loads applied in laboratory conditions and indicated the potential of utilising the technique in a clinical environment by comparing the pressure regions with photoelastic data. The HIPE is expected, after further development, to be able to overcome the most important problems identified in previous pressure measurement studies; interference of sensors on the contact interface and modeling the residual limb using Finite Element Analysis (i.e. unknown tissue properties). It is hoped that this system will eventually become a tool suitable for monitoring the fit of a prosthesis in a clinical environment.
展开▼
