SYSTEM AND METHOD FOR DECENTRALIZED, MULTI-AGENT UNMANNED VEHICLE NAVIGATION AND FORMATION CONTROL

摘要

A system and method for a USV fleet is disclosed that can execute autonomous path-planning in an A2/AD environment, where communication reach back to a base station for instruction or Positional, Navigation, and Timing (PNT) data is not assured. This process has the ability to adapt considering the failure and intermittent connection of the ad-hoc USV-to-USV data network. This process also has the ability to continue operation considering the failure of any or all communications between members and is immediately prepared for reconnection should failures occur.

主权项

1. A method for navigating a fleet of multiple unmanned vehicles from a first starting location to a second ending location via an initial pathway of waypoints, the iterative method for each of the unmanned vehicles comprising:
accumulating sensor information representative of obstacle locations and boundaries of one or more sensed obstacles and vehicle locations of other sensed vehicles in the fleet between the first and second locations; calculating artificial potential field (APF) additive force vectors representative of virtual forces between each of the vehicles and between each vehicle and the sensed obstacles, including calculating when a force vector is zero which is representative of a local minimum; responsive to the local minimum, communicating to the fleet that a new path needs to be generated and generating a new path forward, including calculating one or more new path forward waypoints such that the local minimum is removed because of attraction to the new waypoints by calculating additive force vectors to remove the local minimum and attract the vehicle to the new waypoints around the sensed obstacles and the other sensed vehicles to the second location.