Accessible ground reaction force estimation using insole force sensors without force plates

Ground reaction force (GRF) estimations using instrumented insoles relied on the use of costly force plates in previous works. This paper designed insoles with 15 force sensing resisters and presents an accessible estimation method of vertical GRF (vGRF) using the insole without any cost-prohibitive devices. To determine vGRF from the insole sensor forces, a subject-specific linear regression model was constructed using a least-squares method with a constraint and a bound. The regression matched the sensor forces during single-leg standing (SLS) to a subject's body weight with a linear constraint using data measured while walking. During SLS, standing static and shifting body weight were performed to enhance estimation accuracies. The accuracies of constructed models while walking were evaluated by comparison with the Nintendo Wii Balance Board (WBB) which can measure accurate vGRF compared with force plates. The results for an adult had 8-15 % root mean squared errors (RMSEs) with no significant deviations from previous methods which relied on force plates. From these results, the proposed method was validated as an accessible kinetic gait analysis system.