AI Based Intrusion Detection for Drones

Abstract

Drones are being used more and more in things like watching over places delivering packages checking on disasters and sensing things from away. This means that the amount of information being sent over networks by drones is getting really big. These drone networks need to be able to talk to each other all the time so they can work with ground control stations and cloud servers. This also makes them more open to cyber threats like fake messages extra information being added and people getting in without permission. So, we need safe ways to watch over the traffic on these drone networks. This research is about making a system that uses intelligence to detect when someone is trying to get into the drone network. The system uses a different way to figure out what is going on with the network traffic, including Random Forest Classifier, Gradient Boosting Classifier, Decision Tree and Support Vector Machine. It looks at things like how traffic there is, how different the packet sizes are and what kind of internet control messages are being sent. The system tries to guess what kind of request is being made over the internet and sorts the network traffic to find anything that might be a problem. By learning what the network usually does the system can find things that’re not normal and might mean someone is trying to get in without permission. When we tested the system, we found that the Random Forest and Gradient Boosting classifiers were really good at guessing what was going on and stayed stable when the network was changing. The results showed that using intelligence to detect problems with the drone network made it more secure and worked better. The system helps reduce delays loses information finds bad traffic early and makes sure that information is sent reliably between drones and the control systems. Also using machine learning with drone networks means that the security can adapt to threats. Overall, this way of doing things makes drone networks more able to deal with problems keeps the information safe and can find threats in time which is good, for modern drone applications that work in complicated and busy network environments. Drone networks and drones are important here. We need to make sure drones and drone networks are safe and secure.