Abstract: According to one aspect, uncertainty prediction based deep learning may include receiving, using a memory, a trained neural network policy ? trained based on a first dataset in a first environment, implementing, via a controller, the trained neural network policy ? in a second environment by receiving an input and generating an output y, calculating an uncertainty array U[T] for a time window T, wherein the uncertainty array is indicative of a level of uncertainty associated with an output sample distribution of the output across the time window T based on a temporal divergence, an entropy H, a variational ratio VR, and a standard deviation SD of the output y, and executing, via the controller and one or more systems, an action based on the uncertainty array U[T], such as discontinuing use of the trained neural network policy ?.

Abstract: Disclosed is an online dynamic mutual-observation modeling method for unmanned aerial vehicle (UAV) swarm collaborative navigation, which includes: first performing first-level screening for members according to the number of usable satellites received by a satellite navigation receiver of each member, to determine the role of each member in collaborative navigation at the current time, and then establishing a moving coordinate system with each object member to be assisted as the origin, and calculating coordinates of each candidate reference node; and on this basis, performing second-level screening for the candidate reference nodes according to whether mutual distance measurement can be performed with each object member, to obtain a usable reference member set, and preliminarily establishing a dynamic mutual-observation model; and finally, optimizing the model by means of iterative correction, and conducting a new round of dynamic mutual-observation modeling according to an observation relationship in the U

Abstract: According to one aspect, uncertainty prediction based deep learning may include receiving, using a memory, a trained neural network policy ? trained based on a first dataset in a first environment, implementing, via a controller, the trained neural network policy ? in a second environment by receiving an input and generating an output y, calculating an uncertainty array U[T] for a time window T, wherein the uncertainty array is indicative of a level of uncertainty associated with an output sample distribution of the output across the time window T based on a temporal divergence, an entropy H, a variational ratio VR, and a standard deviation SD of the output y, and executing, via the controller and one or more systems, an action based on the uncertainty array U[T], such as discontinuing use of the trained neural network policy ?.