No abstract

This paper describes a method for the estimation of the 3D ground reaction force (GRF) during human walking using novel nanocomposite piezo-responsive foam (NCPF) sensors. Nine subjects (5 male, 4 female) walked on a force-instrumented treadmill at 1.34 m/s for 120 s each while wearing a shoe that was instrumented with four NCPF sensors. GRF data, measured via the treadmill, and sensor data, measured via the NCPF inserts, were used in a tenfold cross validation process to calibrate a separate model for each individual. The calibration model estimated average anterior-posterior, mediolateral and vertical GRF with mean average errors (MAE) of 6.52 N (2.14%), 4.79 N (6.34%), and 15.4 N (2.15%), respectively. Two additional models were created using the sensor data from all subjects and subject demographics. A tenfold cross validation process for this combined data set resulted in models that estimated average anterior-posterior, mediolateral and vertical GRF with less than 8.16 N (2.41%), 6.63 N (7.37%), and 19.4 N (2.31%) errors, respectively. Intra-subject estimates based on the model had a higher accuracy than inter-subject estimates, likely due to the relatively small subject cohort used in creating the model. The novel NCPF sensors demonstrate the ability to accurately estimate 3D GRF during human movement outside of the traditional biomechanics laboratory setting.

Nanocomposite foam (NCF) is a multifunctional material that can be used to measure impact. Interactions between the flexible polymer matrix and conductive particles dispersed throughout it produce a voltage signal under dynamic strain, which correlates to the magnitude of impact. Though promising in applications requiring both impact sensing and energy absorption, NCF's voltage response has been observed to suffer from significant signal drift. This paper investigates several causes of variance in the response of NCF sensors to consistent impacts. These effects can be classified into three general types: recoverable transient effects (such as those relating to viscoelasticity or capacitive charging), environmental drift (due to humidity and temperature), and permanent signal decay from material degradation. The motivation for the study arises from various potential repeat-impact applications where periodic recalibration of the sensor would be difficult (such as a gait-tracking insole in use for a marathon event). A cyclic drop testing machine was used to apply consistent impacts to NCF, and drift resulting from each factor (in ranges typical of an insole environment) was experimentally isolated. Models representing each factor's contribution to signal drift are presented. Of the factors investigated, humidity and temperature caused the most significant drift, with permanent material degradation accounting for only minor decay in voltage response. Transient effects were also observed, with a characteristic 'warm-up' (or 'charging') time required for the NCF to achieve steady-state; this phenomenon, and the related 'recovery' time for the material to return to its original state, were determined. The resultant data can be leveraged to implement a correction algorithm or other drift-compensating method to retain an NCF sensor's accuracy in both long and short data collection scenarios.

Most mechanical systems produce vibrations as an inherent side effect of operation. Though some vibrations are acceptable in operation, others can cause damage or signal a machine's imminent failure. These vibrations would optimally be monitored in real-time, without human supervision to prevent failure and excessive wear in machinery. This paper explores a new alternative to currently-used machine-monitoring equipment, namely a piezoelectric foam sensor system. These sensors are made of a silicone-based foam embedded with nano-and micro-scale conductive particles. Upon impact, they emit an electric response that is directly correlated with impact energy, with no electrical power input. In the present work, we investigated their utility as self-sensing bushings on machinery. These sensors were found to accurately detect both the amplitude and frequency of typical machine vibrations. The bushings could potentially save time and money over other vibration sensing mechanisms, while simultaneously providing a potential control input that could be utilized for correcting vibrational imbalance.