Obstacle avoidance method and system of UAV formation under time-varying network topology

Abstract

The invention discloses an obstacle avoidance method and system of a UAV (unmanned aerial vehicle) formation under time-varying network topology. The method comprises a step of obtaining a current UAV network topology condition according to position information of each UAV at present, a step of obtaining current UAV global formation information by using a distributed transmission mode according to the current UAV network topology condition, a step of carrying out centralized transmission according to regional information obtained by a sensor of a boundary UAV to calculate a global barrier-free area, a step of carrying out optimal calculation to obtain a target formation according to the current UAV global formation information and the barrier-free area, and a step of distributing a UAV to each target point according to the calculated target formation. According to the method and the system, the overall communication overhead and time cost are reduced, at the same time, a topology change caused by a formation change in the formation process is considered, a slice type topological model is established, different topologies are updated according to a current formation state in different time slots, and thus an algorithm is more close to the actual application. The integrity of the formation is ensured.

Description

technical field [0001] The invention relates to the technical field of multi-UAV formation, in particular to an obstacle avoidance method and system for UAV formation under a time-varying network topology. Background technique [0002] With the rapid development of drone driving technology and the widespread civilian use of drones, people are gradually realizing the benefits of their wide application in different industries. Therefore, with the increasing demand for drones, the demand for multi-machine collaboration to complete tasks is also expanding. The core of multi-machine cooperation to complete the task is the multi-machine formation technology. [0003] At present, the relatively mature and common formation algorithms mainly include the leader-wingman method, the artificial potential field method, the behavior-based method and the virtual structure method. In addition to the above algorithms, there is also the traditional most commonly used PID algorithm. The tradi...

AI Technical Summary

Problems solved by technology

[0005] 1. Most of the algorithms are based on the leader-wingman model. In the research, the leader is not controlled. The cooperative control strategy only works on the wingman. It is a follow-up mechanism. The robustness of the algorithm is poor. When the leader or virtual leader When a problem occurs, the entire formation algorithm cannot work, and there is still a big gap with the real multi-machine formation and cluster formation thinking

[0006] 2. The multi-machine formation system has high requirements for real-time performance. Many control algorithms, such as neural network calculation iterations, are too high and the calculation time is long, which is difficult to implement in engineering. This is also an obstacle to the practical application of many current artificial intelligence algorithms.

[0007] 3. The distributed control of multi-machine formation must consider the changeable problem of the communication link between UAVs. However, many current algorithms are based on ideal communication conditions.

Images