Feature Based Localization and Navigation for Miniature Underwater Vehicles

We propose to design, build, and test a) a low-power instrumentation package for use on small Unmanned Underwater Vehicles, b) autonomous algorithms to detect, characterize, and locate natural and man-made objects on the sea floor, and c) simultaneous localization and navigation algorithms, for use along optimally designed paths, to produce extremely accurate environmental maps that will be used to reduce navigation errors.Our primary sensor is a proprietary side-scan sonar that we sell commercially and the secondary sensors are a Doppler Velocity Log and an Inertial Navigation Sensor.The strengths of our instrumentation approach are a) the way that these 3 sensors are integrated to reduce mutual interference, improve performance by reducing size, weight, and power, compared to separate systems, and b) the way that the sensors adapt their parameters to exploit changing physical conditions.The strengths of our algorithms are a)the way that we solve the multiple objective optimization problem, and b) how we navigate from a known location to a distant operational area, such that we simultaneously maximize time-on station (to conduct a mission) and minimize navigation errors when returning to the starting point.