Low-altitude flight safety management method under multi-source data monitoring

Abstract

The invention discloses a low-altitude flight safety management method under multi-source data monitoring, and relates to the technical field of low-altitude flight safety management, and the method comprises the steps: employing a target detection algorithm to extract a flight-related object, and obtaining the spatial position parameter and geometric feature parameter of the flight-related object based on a three-dimensional point cloud reconstruction technology, the method comprises the following steps: dividing a low-altitude airspace into three-dimensional grids, independently calculating a risk index for each grid unit, carrying out risk level evaluation on each airspace grid through a set multi-dimensional threshold judgment model, constructing a feasible path set according to the current position and the target position of an aircraft in combination with the risk levels of the airspace grids, and carrying out risk assessment on the feasible path set. Modeling a dynamic interaction relationship between the aircrafts by adopting a graph neural network, modeling the dynamic interaction relationship between the aircrafts, and predicting a possible interaction conflict risk; based on the predicted possible interaction conflict risk between the aircrafts, static nonlinear control and dynamic linear control are taken as theoretical references, flight area decision control is performed on a low-altitude airspace, and flight decisions of the aircrafts are adjusted in real time in combination with risk levels, obstacle distribution and dynamic states of the aircrafts; and performing path optimization by using a Q-learning reinforcement learning algorithm to generate a flight path considering both global optimality and local safety. The technical problems that in the prior art, multi-modal data cannot be subjected to high-precision space-time alignment, and real-time performance and safety of flight target recognition and path planning are difficult to consider at the same time are solved, and the technical effects of aircraft efficient sensing and intelligent path generation in a complex airspace are achieved.

Images