Collaborative communication and control for drone swarm path planning

Abstract

The importance of swarm drones is growing in various fields such as industry, airborne service, and national defense. In the industry, drones are being actively used for work in areas that are difficult to reach or work directly, such as bridges and facility inspections and map making, and efficiency increases such as improved work speed and precision can be expected when using multiple drones. In defense, the use of multiple drones for various applications such as surveillance, reconnaissance, strike, electronic warfare, and communication relay is expected to enhance operational effectiveness and build low-cost power. In order to make multi-drone a reality, one person must be able to control the multi-drone extremely, which is like in a game, when you command all the drones to move to a specific point, the drones automatically avoid collisions and plan a multi-agent route to move to the destination. This should be a prerequisite. And it is required to build a network that actively supports this. In other words, a dedicated network is required to meet mission characteristics-based demand delay for transmission and reception of mission and status information between ground operators and drones considering the scale scalability of swarm drones, and information transmission and reception between drones for collision avoidance and collaboration.

In multi-agent path planning, there are methods to detect the location of adjacent drones through vision, radar, lidar sensors and communication and to estimate the probability of collision to find a path to the destination while avoiding collision in a geometric or force field method. Considering the cost, the use of expensive sensors such as lidar and radar may be limited in order for swarm drones to be effective, so it is reasonable to use communication as a low-cost method to detect the location of adjacent drones. However, unlike most static sensors, the communication topology of drones changes rapidly due to high-speed movement in the air, so high-speed update of information is essential for accurate sensing of neighboring drones. The challenge is that high-speed updates cause extreme communication loads with the increase in the number of drones. Under limited communication resources, an increase in load raises the communication failure rate, which leads to the failure of detecting neighbors or obstacles, resulting in fatal drone crashes. Therefore, in order to prevent this, not only efforts to increase the communication success rate in traffic expansion, but also communication and route planning are organically connected, so that it is required to establish a cooperative system that enables the flight of swarm drones even in low-speed or poor communication environments. Therefore, we propose a method for improving communication performance for stable flight of swarm drones and improving flight performance through collaboration between communication and route planning.

First, the transmission of location information, which is broadcast with a radio wave arrival distance and transmission period greater than the minimum distance between drones for collision prevention, is the traffic that has the greatest impact on the network load. We propose a method to reduce media access delay.

Second, as another method for improving communication performance, we propose a clustering technique that minimizes additional information in the transmission of status information that each drone periodically transmits to the ground control center. Although the amount of status information is small compared to the previous location information, it causes wide-area interference due to a long communication distance and causes packet collisions due to location information transmission with a relatively short transmission distance and carrier detection asymmetry. Therefore, a clustering technique that reduces the number of wide-area transmissions as much as the number of drones is required.

The proposed method is compared with the existing method through the OPNET-based route planning/communication integrated simulator and the real environment swarm drone flight test. Considering the situation of simultaneous operation of multiple missiles/unmanned aerial vehicles in national defense, it was confirmed that the proposed method improves the speed to the target point by up to 28\% and guarantees no collision in the path for the maximum speed of sound movement. To achieve this, it was confirmed that the proposed method optimizes the force field constant in the artificial potential field (APF)-based path planning and the transmission period parameter in the WiFi-based communication method. When two or more drones come close within a certain distance, the collision risk is detected by increasing the information exchange speed, and the maximum speed including collision avoidance is determined by determining the corresponding speed.

In conclusion, the stability and application of swarm drone operation is required by ensuring non-collision by adapting to changes in the drone's flight environment and maximizing the speed to the destination through the communication/control collaboration technique for route planning of multiple drones proposed in this way. performance could be satisfied.