Artificial Intelligence/Machine Learning to Improve Maneuver of Robotic/Autonomous Systems

Autonomous vehicles are being used for many applications varying from small indoor robots, to large-scale mining trucks, and military vehicles operating in fast-changing environments. Global position systems (GPS) have traditionally been used for vehicle positioning. GPS works well in open environments, but suffers from significant degradation due to terrain, certain operating conditions, and even interference and jamming. Vehicles operating in high-risk scenarios, whether the battlefield or a time-sensitive mine operation or farming harvest, need to always know where they are in order to properly perform path control, particularly in dynamic environments. In these cases, GPS alone will not suffice for vehicle state estimation. Under the previous Phase I contract, Autonomous Solutions, Inc. (ASI), in conjunction with the Georgia Institute of Technology, have developed both GPS-denied and high-speed dynamic control algorithms to mitigate the effects of degrading GPS and to improve vehicle control, respectively. The technology demonstrated shows promise in its usefulness to solve many problematic situations currently expressed by ASI customers in non-ideal environments. The continued development of this GPS-denied work will prove beneficial to ASI customers operating in deep open-pit or underground mines, orchards with impenetrable canopies, and yet others operating at high-latitudes where GPS is unreliable.