The Math Model of Drone Behavior in the Hive, Providing Algorithmic Architecture

Introduction

Nowadays there are many types of drones on the market. One can purchase not just a ready to use drone but also different types of the component parts used in drones’ modeling. The types of the drones also vary a lot: from the smallest ones (usually used for gaming, video streaming or military purposes) to the big ones used for transporting different kind of cargo and even people (such as Ehang-184). It’s possible to purchase even the controllers and the software which can already implement the basic operations necessary for the drone, such as different kind of movements, etc. So, a developer having the SDK can work on more complex topics such as managing the hive of drones, applying AI, etc.

A developer will have to implement many different functionalities and apply lots of technologies working on such the projects.

For example, below are listed some modules which will be necessary to develop for such the work:

Computer vision, which is usually based on OpenCV or similar frameworks.

Voice recognition system which is based on transforming of voice into text and then a bot system which would be able to properly process the text and select the proper action from the TODO list. There are lots of different frameworks and libraries for this purpose, for example Microsoft bot. Some advanced system are based on different kind of neural networks and different kind of frameworks like TensorFlow, etc.

The module responsible for different type of calculation of the flight path and trajectories, which would depend not on the most short path but first of all on the physical abilities of a drone, taking into consideration the angle of rotation a drone can take, the speed, the weight of the drone, the type of the drone (cargo, taxi, etc.), etc. As an example: it’s not the proper path when we consider a taxi drone carrying people and the path of the flight has a rotation of 90 degrees and high speed.

• •

  The module for checking the positions and the status of the other objects (not only the drones) in the area and a predictable system which could send urgent signals to the drones to avoid their collision.

During the research and the work of the author, the author developed multi-layered architecture and software for management of the hive of drones.

The system manages not only the drones but the whole drone’s traffic over some territory. This could be compared with management of the regular road traffic.

The author developed several architectures for such the task, one of them is related to management of the hive and not related to the drone’s traffic management. Such the system could be used in the military sphere. Another system manages the traffic and the drones and it could be applied in civil life.

In this article we will consider the first type of the system, for managing the hive of the drones.