Unmanned aerial vehicle path planning method based on artificial fish schools

Abstract

The invention provides an unmanned aerial vehicle path planning method based on artificial fish schools. According to the invention, a visual field and a step length are improved on the basis of artificial fish schools, so that the artificial fish schools have self-adaptive capacity, a congestion degree factor is ignored at a low probability in the early stage of solving, local optimum is avoided, and the congestion degree factor is ignored at a high probability in the later stage so as to prevent fish schools from wandering around the global optimum and increase the convergence speed; in order to avoid early maturing caused by blindness of rear-end collision, abandoning operation is adopted to carry out rear-end collision on the worst individual with a certain probability, and local optimum is jumped out, so that the situation that individual fish schools tend to local optimum and accordingly useless search is carried out is avoided while the diversity of the groups is expanded; and through experimental simulation, the method is used for simulating unmanned aerial vehicle path planning, the effect far exceeds that of a basic artificial fish swarm algorithm, and the accuracy of path planning is greatly improved in the aspect of searching the most significant solution.

Description

technical field [0001] The invention relates to the technical field of unmanned aerial vehicles, in particular to a path planning method for unmanned aerial vehicles based on artificial fish swarms. Background technique [0002] At present, UAVs are generous in various fields and are widely used in investigation, delivery, rescue and military fields. Since UAVs travel in the airspace and overcome most constraints of geographical conditions, they have extremely high advantages over traditional tools. Then in practical applications, in addition to visible obstacles, such as tall buildings and mountains, which can be observed by the camera on the drone and avoided artificially or intelligently, similar electromagnetic interference, signal interference or facilities in the field that may shoot down drones, etc. Unable to observe intuitively, it often relies on inference to divide the approximate range to form inaccessible areas. How to make drones avoid these invisible areas and...

AI Technical Summary

Problems solved by technology

[0004] The artificial fish swarm algorithm does not rely too much on the function gradient, and only determines whether the problem is solved by comparing the fitness value. It has a certain degree of randomness in the optimization process, and has the ability to avoid the local optimum and obtain the global optimal solution. However, since most of the parameters need to be manually formulated during the algorithm initialization, and run through the algorithm, the change of a single parameter can affect the entire solution process

The foraging behavior is the driving force for finding the optimal solution. The preset number of attempts determines the search characteristics of the algorithm. With the continuous iteration, the number of attempts set at the initial stage cannot be applied to the later optimization. In addition, random exploration has blindness. There is an individual fish hovering at a certain local optimal point, which increases the calculation cost of the algorithm operation; in the later stage of the clustering behavior, the fitness value of the individual fish tends to be the same, due to the existence of the crowding factor, when the individual fish gradually moves towards When the optimal solution approaches, it is often excluded, and the optimal solution cannot be further obtained; the tail-chasing behavior adopts the greedy principle, and only approaches the better solution within the field of vision while ignoring the overall situation. In the later stage of the algorithm, the tail-chasing is blind. If the tail-chasing target is local Optimum will lead to a large number of artificial fish, not only easy to fall into it, lack of ability to jump out of the local optimum, but also make a large number of individual fish do useless work, wasting the computing power of the system

Images