Comparative testing of piezoelectric and printed strain sensors in characterization of chewing

Results of recent research suggest that there may be a relationship between the eating rate and the total energy intake in a meal. The chewing rate is an indicator of the eating rate that may be measured by a sensor. A number of wearable solutions have been presented for the automatic detection of chewing, but little work has been done on counting chews automatically. With recent developments in printing technologies, it is possible to draw or print application specific sensors. This paper provides a comparison between an off the shelf piezoelectric strain sensor and a plotter drawn strain sensor for quantifying the number of chews for several food items. Piezoelectric strain sensor and plotter drawn strain sensors were able to achieve absolute mean error rates of 8.09 ± 7.16% and 8.26 ± 7.51% respectively for estimating the number of chew counts. This shows that a plotter drawn sensor can achieve similar performance while potentially providing an easily reconfigurable solution.