Multi-Inlet 2-D Water Flow Vector Sensor Using a Sealed Incompressible Liquid and Neural Network for Marine Biologging

This study designed and demonstrated a 2-D multiinlet water flow sensor to address challenges associated with measuring the swimming speeds of marine animals owing to size limits and harsh marine environmental conditions, such as clogging by marine debris. The sensor comprises a spherical housing equipped with three pairs of inlets on opposite sides, each connected to a differential pressure sensor element through a tube filled with an incompressible liquid. Notably, a sealed film placed over each inlet prevents marine debris from intruding into the sensor. The differential pressure data collected by the three sensor elements are used to determine 2-D water flow speed and angle using a neural network (NN) trained on water tunnel experiments. The results of the experiments conducted in this study demonstrated that the developed sensor could accurately detect the water flow speed and angle within a speed range of 0.6–2.0 m/s, which is typical of green turtles. The proposed sensor is, therefore, expected to realize improved biologging of marine animal behaviors.