Lateral-Line-Inspired Sensor Arrays for Navigation and Object Identification

Abstract

AbstractThe lateral line is a critical component of fish sensory systems, found to affect numerous aspects of behavior, including maneuvering in complex fluid environments with poor visibility. This sensory organ has no analog in modern ocean vehicles, despite its utility and ubiquity in nature, and could fill the gap left by sonar and vision systems in turbid, cluttered environments.To emulate the lateral line and characterize its object-tracking and shape recognition capabilities, a linear array of pressure sensors is used along with analytic models of the fluid in order to determine position, shape, and size of various objects in both passive and active sensing schemes. We find that based on pressure information, tracking a moving cylinder can be effectively achieved via a particle filter. Using principal component analysis, we are also able to reliably distinguish between cylinders of different cross section and identify the critical flow signature information that leads to the shape identification. In a second application, we employ pressure measurements on an artificial fish and an unscented Kalman filter to successfully identify the shape of an arbitrary static cylinder.Based on the experiments, we conclude that a linear pressure sensor array for identifying small objects should have a sensor-to-sensor spacing of less than 0.03 (relative to the length of the sensing body) and resolve pressure differences of at least 10 Pa. These criteria are used in the development of an artificial lateral line adaptable to the curved hull of an underwater vehicle, employing conductive polymer technologies to form a flexible array of small pressure sensors.