Virtual reality application for real-time pedalling cadence estimation based on hip ROM tracking with inertial sensors: a pilot study

Abstract

Virtual reality (VR) applications on rehabilitation a home-base exercise experiences have boomed in the last decade. This is mainly because their entertainment capacity creates a sense of immersion in the users, which enhances adherence to their use. In addition, offering body-related visual feedback is a proven approach to the physical training towards a goal. Recent literature showed the exercise of pedalling has the potential to provide a high number of flexion and extension repetitions of the lower limb in reasonable therapeutic time periods to improve muscle activity, strength and balance in elders, but also motor improvements in patients with neurological injuries. The objective of this work is to present a low-cost wireless application in virtual reality (VR) for pedalling exercises. The platform developed consists of a VR headset and an inertial measurement unit (IMU). The VR headset processes the kinematic information of the IMU to estimate the cadence of the pedalling, while the IMU sensor tracks the angle of hip flexion/extension movement of the user. In order to confirm the suitability of this cadence estimation system, our approach is confronted with a cycling platform developed and validated in a previous study. In the present study, we carried out two repeated sessions with 13 subjects at 3 set speeds: slow (30 rpm), medium (60 rpm) and fast (90 rpm). The Spearman’s correlation (PC) between both systems for the 3 speeds and sessions shows high correlation values for low and medium speeds and moderate correlation for high speed. The SEM results for each system show low measurement error (about 1 cycle) for both systems at every target speed, except for the virtual cycling platform at the highest speed (SEM of VCP at 90 rpm = 3.24 cycles). The repeatability analysis based on ICC (3, 1) absolute agreement shows consistency in all measurements for both systems at high speed and also reflects the irregularity in measurements at low and medium speeds, where participants were less stable during testing due to entertainment from the VR system. All in all, it is concluded the validity of the cadence estimation system for pedalling exercises with low intensity. This development allows us to control the virtual environment by adapting the visual stimulus to cycling cadence. The proposed system can generate sensitive inputs to influence the user’s pedalling cadence.