In this report, we investigate the accuracy of RSI (Reactive Strength Index) measured by sensor insoles.
The experiments were conducted with participants wearing sensor insoles (Plantiga Technologies, 2020) and performing consecutive countermovement jumps on a Dual force plate platform (Kistler). Data was collected and processed for both the sensors and force plates. After processing, RSI was computed from both and the results were compared.
Linear Regression analysis showed a coefficient of determination (r-squared) value of 0.985 and a slope of 1.08. These results show that Plantiga sensor insoles can provide a reliable and accurate way of measuring RSI while providing the benefit of being lightweight and portable.
The Reactive Strength Index (RSI) is a way to assess an individual’s strength and power using plyometric exercises. It reflects an athlete’s ability to rapidly change from an eccentric (braking or deceleration) movement to a concentric (propulsive) movement [1].
RSI is trainable [2] and in sports that require fast vertical jumping ability and changes of direction, it is quite valuable. It is also sensitive to an accumulation in training load (workload), recovery and tapering/peaking [3, 4]. Characterizing these natural responses to training matter greatly to coaches and trainers because ultimately the goal is to prescribe training in order to improve performance.
A common method to measure Reactive Strength Index is derived from the measurement of flight time and ground contact time. Mathematically, it is calculated as
Plantiga insoles measure these times with milliseconds accuracy via Signal Processing and Machine Learning Techniques. The location of the motion tracker (IMU) in the insole allows measurements to be made as close to the ground as possible.
Jump data was collected from 13 individuals while a total number of 60 jumps were analyzed. Each participant was asked to wear the insoles and perform 2 trials of consecutive countermovement jumps. They were instructed to land and take-off in a similar way while keeping their hands on their hips. For comparison, a dual force plate platform (Kistler) was used to compute flight time and contact time which was then used to compute the reference RSI.
The Force Plate sampled the signal at a much higher rate of 1000 Hz compared to the insole/imu sampling rate of 500 Hz. Therefore, once the two signals were synced, the force plate data was downsampled to match the insole sampling frequency of 500 Hz. Results Linear regression analysis was performed on RSI computed from both the Reference Force Plates and Sensor Insoles. A general linear relationship was found between the two. The coefficient of determination was found to be 0.985 with a slope of 1.080respectively. This is shown in Figure 1.
Figure 1: RSI Regression
Figure 1 shows linear regression analysis between Insole RSI and Force Plate RSI.
In this report, RSI values computed by sensor insoles were compared to the reference values provided by the force plate platform. The results show a linear relationship between the two with r-squared values of 0.985 and a slope of 1.08. This shows excellent correlation. With this, we can conclude that the RSI values predicted by the insoles are reliable and accurate. Further, the added benefit of portability makes these insoles highly applicable to high performance sports and rehabilitation.