An artificial intelligence approach for measurement and monitoring of pressure at the residual limb/socket interface - A clinical study

Amali, Ramin; Noroozi, Siamak; Vinney, John; Sewell, Philip; Andrews, Stephen

Abstract

A good-fitting prosthetic socket requires the pressure between the stump and socket to be distributed to ensure the load is carried by pressure tolerant regions of the limb. Transducers and Finite Element Analysis have been utilised to measure and monitor these pressures. However, it has been recognised that both techniques have limitations, making them impractical for everyday clinical use to aid the prosthetist in the socket fitting process. This paper details the design of a Hybrid Inverse Problem Engine (HIPE) which combines Artificial Intelligence (AI) and experimental/numerical data to create a less invasive and passive approach to develop a practical clinical tool for predicting the pressure distribution at the limb/socket interface. Testing and validation of the HIPE under laboratory conditions showed that the technique was able to predict the location and magnitude of pressures applied manually to the socket. A comparison of the predicted pressure distribution found using the HIPE, at the limb/socket interface of a patient in a clinical environment with photoelastic data of the actual pressure distribution, further indicated the technique's potential benefits. It is hoped that the HIPE will eventually become a general tool suitable for monitoring the fit of a prosthesis in a clinical environment.

• Journal Article Type: Article
• Publication Date: Jul 1, 2008
• Deposit Date: Dec 11, 2012
• Journal: Insight: Non-Destructive Testing and Condition Monitoring
• Print ISSN: 1354-2575
• Publisher: British Institute of Non-destructive Testing
• Peer Reviewed: Not Peer Reviewed
• Volume: 50
• Issue: 7
• Pages: 374-383
• DOI: https://doi.org/10.1784/insi.2008.50.7.374
• Keywords: artificial intelligence, residual limb/socket interface pressure, ann, prosthetics, hybrid inverse problem engine
• Public URL: https://uwe-repository.worktribe.com/output/1010875
• Publisher URL: http://dx.doi.org/10.1784/insi.2008.50.7.374
• Contract Date: Nov 15, 2016