End-side collaborative multi-unmanned aerial vehicle autonomous navigation method

Abstract

The invention relates to an end-side collaborative multi-unmanned aerial vehicle autonomous navigation method. The method comprises the following steps of S1, carrying out autonomous navigation on an unmanned aerial vehicle by adopting a convolutional neural network, S2, quantitatively analyzing the influence of end-to-end time delay on navigation, S3, adjusting the resolution of the input image through a spatial pyramid pooling method, and reducing the calculation time delay, S4, defining a single unmanned aerial vehicle navigation optimization problem, defining a state space of reinforcement learning, defining an action space of reinforcement learning, and defining a reward of reinforcement learning, S5, in a multi-unmanned-aerial-vehicle scene, estimating the probability that each unmanned aerial vehicle is unloaded to the edge server, according to the unloading probability, preliminarily allocating computing resources to each unmanned aerial vehicle, defining an upper limit and a lower limit of computing resources of the unmanned aerial vehicle to ensure fairness, adjusting a computing resource allocation scheme to enable the resources obtained by each unmanned aerial vehicle to be smaller than a predefined upper limit, and adjusting a computing resource allocation scheme to enable the resources obtained by each unmanned aerial vehicle to be greater than a predefined lower limit. The method is higher in accuracy and robustness.

Description

technical field [0001] The invention relates to the technical field of autonomous navigation of unmanned aerial vehicles, and more specifically, relates to a multi-unmanned aerial vehicle autonomous navigation method with end-edge coordination. Background technique [0002] UAV is a terminal computing device. In order to solve the problem of insufficient computing power of terminal devices, the traditional method is to offload computing tasks to the cloud. However, the bandwidth between the terminal device and the cloud is often unstable, which will increase the transmission delay. In order to solve this problem, researchers have proposed the concept of edge computing, which uses servers deployed at the edge of the network to cooperate with terminal devices to perform computing-intensive tasks, which is called end-edge collaboration. The edge server refers to a type of server that is deployed on the edge of the backbone network and geographically close to users. The commun...

AI Technical Summary

Problems solved by technology

[0004] However, when using the "model bag" method, different models need to be dynamically switched online. The switching process will introduce model loading and initialization delays, which will drag down system performance.

Moreover, multiple models of different sizes require resident memory. For terminal devices with limited memory space, the additional memory usage is unacceptable.

Moreover, the above-mentioned system models the relationship between all decision variables and optimization objectives with a mathematical model, and the mathematical relationship between variables is fitted through empirical observation data. This fitted relationship is inaccurate, and the online fitting These relationships create additional system overhead

Images