DISTRIBUTED DENIAL OF SERVICE (DDoS) ATTACK ON UNMANNED AERIAL VEHICLE

Abstract

The Unmanned Aerial Vehicle aka Drones are the new devices which are evolving at great pace from last decade. Before that the drones found the application in military warfare only. But as the technology is evolving it is making impact on this device as well. Now days these drones are also called as Internet of Flying Things (IoFT) as they are getting inspiration from Internet of Things (IoT) devices. As these IoT devices are getting adapted in modern world these IoFT are also finding the application in various sectors like agriculture, medical emergencies, disaster prone areas, searching and surveillance industry, delivery industry etc. These IoFT devices can make use of internet for communication purposes between them as well as to ground station where they are controlled. So as these devices make use of internet it is also capable of getting attacked by different types of cyber-attacks. Distributed Denial of Service or DDoS attack is one the cyber-attack which can be disastrous in nature if not handled properly. This attack as if do nothing to device but it will eat all the resources of the network of drones which we can also refer to Internet of Drones (IoD). So as the resources like bandwidth of communication channel is full it can cause the drones to get blind such that it will not receive the correct signals as its bandwidth is full, So it can result in harm for public safety as the drone can crash and also for stealing the drone and its information. So DDoS detection for drones is one of the important task. Machine learning is evolving technique which is useful for many tasks and one of the task is prediction analysis. So in order to detect the DDoS attack on UAV machine learning models can be used. So this report discuss the experiment which is used in order to get our dataset by making a setup where drone is been DDoS attacked and data has been collected. Then this dataset is used to predict the attack and also it represent the feature extraction technique in order to improve the accuracies of the machine learning models and making them more efficient to use.