Muñoz Organero, MarioLittlewood, ChrisParker, JackPowell, LaurenGrindell, CherylMawson, Sue2018-03-222018-03-222017-06-15IEEE Sensors Journal, (2017), 17(12), 3909-3920.1530-437Xhttps://hdl.handle.net/10016/26538Insole pressure sensors capture the different forces exercised over the different parts of the sole when performing tasks standing up. Using data analysis and machine learning techniques, common patterns and strategies from different users to execute different tasks can be extracted. In this paper, we present the evaluation results of the impact that clinically diagnosed osteoarthritis of the knee at early stages has on insole pressure sensors while walking at normal speeds focusing on the effects caused at points, where knee forces tend to peak for normal users. From the different parts of the foot affected at high knee force moments, the forefoot pressure distribution and the heel to forefoot weight reallocation strategies have shown to provide better correlations with the user's perceived pain in the knee for OA users with mild knee pain. This paper shows how the time differences and variabilities from two sensors located in the metatarsal zone while walking provide a simple mechanism to detect different strategies used by users suffering OA of the knee from control users with no knee pain. The weight dynamic reallocation at the midfoot, when moving forward from heel to forefoot, has also shown to positively correlate with the perceived knee pain. The major asymmetries between pressure patterns in both feet while walking at normal speeds are also captured. Based on the described features, automatic evaluation self-management rehabilitation tools could be implemented to continuously monitor and provide personalized feedback for OA patients with mild knee pain to facilitate user adherence to individualized OA rehabilitation.11application/pdfeng© 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.Insole pressure sensorsOsteoarthritisMachine learningClassificationresearch articleMedicinaTelecomunicacioneshttps://doi.org/10.1109/JSEN.2017.26963033909123920IEEE Sensors Journal17AR/0000020706