Verifying IMU Suitability for Recognition of Freshwater Mussel Behaviors

Abstract

Freshwater mussels are both an important and threatened part of our ecosystem. They are extremely efficient in cleaning natural water resources, yet surprisingly little is known about the behaviors of these animals, making threat mitigation difficult. As activity recognition based on the use of accelerometers is known to be effective in humans and other animals, this work applies such methods to the study of freshwater mussels in order to aid further scientific study. In this paper, we present results from a laboratory experiment where three inertial measurement units (IMU) were attached to the shell of mussels for approximately 24 to 48 hours. Over this period, we observed several interesting behaviors and verified the sensor arrangement can differentiate active and inactive periods as well as quantify gape frequency, duration, and intensity. Our future work is to develop a complete mussel sensor network where multiple mussels can be monitored in situ for an extended period of time. By better understanding the behavior of these animals, we aim to develop new ways to quantify ecosystem health and more effective conservation strategies.