Flow Vision for Autonomous Underwater Vehicles via an Artificial Lateral Line

Most fish have the capability of sensing flows and nearby movements even in dark or murky conditions by using the lateral line organs. This enables them to perform a variety of underwater activities, such as localizing prey, avoiding predators, navigating in narrow spaces, and schooling. To emulate this capability for Autonomous Underwater Vehicles, we developed an artificial lateral line using an array of Micro-Electro-Mechanical-Systems (MEMS) flow sensors. The signals collected via the artificial lateral line are then processed by an adaptive beamforming algorithm developed from Capon’s method. The system produces 3D images of source locations for different hydrodynamic activities, including the vibration of a dipole source and the movement of a tailflicking crayfish. A self-calibration algorithm provides the capability of self-adaptation to different environments. Lastly, we give a Cramer-Rao bound on the theoretical performance limit which is consistent with experimental results.