Robotic operation of uncrewed sea surface vehicles (USVs) offers a useful analogue for the planetary rovers and other spacecraft of the future—they operate in largely unstructured environments, relying on onboard sensing and decision making, with only sparse supervisory control input from distant human operators. In this task, JPL Robotics is developing a range of autonomy and perception technologies, based on the CARACaS (Control Architecture for Robotic Agent Command and Sensing) autonomy architecture; the primary thrusts are:
Hazard Avoidance: USVs require robust motion planning algorithms to avoid sea surface hazards while complying with the maritime rules of the road (known as COLREGS) and accounting for special circumstances such as towed sensor arrays.
Autonomous Swarm Operations: To enable joint missions employing a team of vehicles (such as harbor patrol or vessel escort), JPL develops software for mission planning, task allocation, and cooperative behaviors within the CARACaS architecture.
Visual Object Detection and Analysis: JPL’s computer vision expertise continues to extend the situational awareness capabilities of maritime vehicles, using novel extreme-range stereo vision technology, machine learning to classify vessel types for COLREGS compliance, and analyzing distant vessels using multiple perspectives from swarms of USVs.
People on this Task