Design of Swarm Technology using Image Processing and Convolution Neural Networks

The proposed research work focuses on the design of swarm technology using image processing and convolution neural networks (CNN). The aim and objectives of the project deal with coordination of communication and exchange of information among multiple drones. Initially a swarm is defined with 3 drones and implemented with Airsim Library. To understand the behavior of drones, dynamics of unmanned aerial vehicles were studied. The various hardware components and sensors (Infrared/GPS) were analyzed using Airsim Library. The mathematical modeling of UAVs and their control commands was implemented using Unreal Engine, Python, and the Spyder IDE. A swarm of three drones was tested, each integrated with image processing software and IR/vision systems. Convolutional Neural Networks (CNNs) were employed to enhance the visual processing capabilities.