Design and Characterization of Shoe Embedded Pressure Sensors for Gait Analysis and Rehabilitation

Abstract

In clinical gait therapy, the quality of gait analysis is critical for developing a training plan and monitoring patient progress. Ground contact forces (GCFs) are often recorded to estimate joint torques which can quantify a patient’s needs and strength development. They are also useful in designing and controlling rehabilitative and assistive devices. In clinical gait analysis, force plates are used to measure GCFs objectively and precisely [1, 2]. Currently, forces sensitive resistors (FSR) are often used as a mobile platform to measure GCFs. FSR based platforms exhibit considerable hysteresis and have low durability, some requiring replacement after only 5-hour long uses. As an alternative to FSR, a pair of sensor-embedded shoes (smart shoes) relying on air pressure sensors has been presented in previous research [3]. Some details regarding the precise characteristics of the sensing abilities were unknown, though, generating unanticipated errors during use. In this paper, the sensing units of wireless smart shoes are characterized and tested to verify their capability to provide real-time and accurate GCF measurements. For the prototype, silicon tubes were sealed on one end, wound into coils, secured to the underside of the shoe’s insole at four points of interest (heel, toe, the first and fourth metatarsophalangeal joint) routed outside the shoe, and their open ends are connected to air pressure sensors as shown in Fig. 1(a). The pressure sensors were placed on a circuit board along with a battery and microcontroller responsible for reading sensor outputs and wirelessly communicating data to a nearby device, as shown in Fig. 1(c). The sensing unit on the lateral side of the shoe is 1.2″ × 1.3″ × 3.95″. A series of calibration tests were first performed on the tube-insole subsystem in isolation to test linearity, repeatability, and hysteresis. Then practical experiments were performed on a healthy subject to determine the accuracy of GCF measurement. A previously presented hysteresis filter was implemented in practical testing [4].