Research On Multicopter Detection Of GPS Hijacking Attacks And Resistance For Denial Of Service Attacks

Unmanned aerial vehicles(UAVs)are increasingly used in military and civil fields,such as strategic reconnaissance,electronic warfare,military strikes,precision agriculture,aerial remote sensing,and product delivery,is contributing to the growth of the UAV market.MarketsandMarketsTM estimates the UAV market size will increase to USD 45.8 billion by 2025,from USD 19.3 billion in 2019,including USD 26.8 billion of the military UAVs' market.Since UAVs have no human operators onboard and the functions of navigation and flight attitude control heavily rely on the real-time transmission of satellite signals,control data and real sensing data of onboard sensors,they have many potential security risks which are extremely vulnerable to attacks,such as Global Positioning System(GPS)hijacking attacks,Denial of Service attacks,false data injection attacks,Trojan horses,multi-protocol attacks,cross-layer attacks,eavesdropping attacks,etc.However,for GPS hijacking attacks,there are few detection methods that do not need additional hardware or are suitable for UAVs with low processing chips.At present,there are few resistance techniques to guarantee the normal flight of UAV after being attacked by denial of service attack.Therefore,this dissertation aims at the detection of GPS hijacking and the resilience of DoS attacks respectively.UAVs generally use GPS and Inertial Navigation System to navigate themselves,while the GPS positioning process depends on the received satellite information on mat-ter it is fake or real.Attackers could easily launch GPS hijacking attacks by forging or forwarding GPS data,thus they can illegally obtaining sensitive data on UAVs after they crash or land at unplanned locations.Therefore,it is extremely urgent to work on how to detect GPS hijacking attacks.For this kind of attack,this dissertation proposes three methods to detect such attacks.On the other hand,since the data published or subscribed by both onboard sensors and actuators needs to be received in real-time,so the UAV can immediately adjust its flight attitude according to the surrounding environment.The attacker can make UAVs unable to receive the above information in real-time through Denial of Service attack,causing the UAV to crash.Aiming at this kind of attack,this dissertation proposes a resilient framework to defense multicopter denial of service at-tacks.In addition,this dissertation focuses on the research of multicopter which is an important category of UAV.The contributions can be summarized as follows:(1)This dissertation proposes a detection method based on onboard Inertial Mea-surement Unit(i.e.,gyroscopes and accelerometers)and GPS.Using the angular velocity measured by gyroscopes and the position reported by GPS to estimate the acceleration in body-fixed coordination,and such accelerations are compared with the data published by accelerometers.In this way,some accumulated errors caused by integration can be avoided,and this method could achieve high detection correctness.(2)Since when the multicopter is not hijacked and the flight path is a curve,the ac-curacy rate of the detection result in the above method is not high enough,this dissertation proposes the second GPS hijacking detection method,which is only based on gyroscopes and GPS.Ideally,GPS data and the angular velocities measured by gyroscopes can be used to estimate the acceleration of a multicopter,which can be further compared with the measurement of the accelerometer to detect whether a multicopter has been hijacked.However,the detection results may not always be correct due to some calculation and measurement errors,especially in curve trajectory situations when no hijacking occurs.Therefore,we propose a novel and simple method to detect hijacking only based on gy-roscopes measurements and GPS data,without using any accelerometer in the detection procedure.The computational complexity of our method is low,which is suitable to be implemented in the multicopters with micro-controllers.On the other hand,the proposed method does not rely on any accelerometer to detect attacks,which means it receives less information in the detection procedure and may decline the accuracy of the results in some specific situations.While the previous method can compensate for this method,the high detection results also can be guaranteed by using the above two methods together.(3)Since the detection accuracy of GPS hijacking for the above two methods is heavily dependent on the selection of the detection threshold,this dissertation proposes a two-step GA-XGBoost method to detect GPS spoofing attacks.Firstly,it uses flight logs to pre-training the model offboard.Then,in order to achieve higher detection results on different multicopters,when the model is loaded on for the first time,it will continue to train the model onboard before prediction.In addition,this method uses genetic algo-rithm to adjust the training parameters offboard so as to quickly and effectively find the appropriate parameters' values and improve the prediction accuracy.(4)This dissertation proposes a software framework that offers DoS attack-resilient control for real-time multicopter systems using containers.The framework provides de-fense mechanisms for three critical system resources:CPU,memory,and communication channel.This method restricts attacker's access to CPU core set and utilization.Memory bandwidth throttling limits attacker's memory usage.By simulating sensors and drivers in the container,a security monitor constantly checks DoS attacks over communication channels.Upon the detection of a security rule violation,the framework switches to the safety controller to mitigate the attack.In summary,this dissertation studies the multicopter,especially the detection of GPS hijacking attacks and the defense of denial of service attacks.The results of this dissertation contribute to theoretical foundations,also provides practical insights for the multicopter.