Alireza Abbaspour

Neural adaptive observer-based sensor and actuator fault detection in nonlinear systems: Application in UAV

A new online detection strategy is developed to detect faults in sensors and actuators of unmanned aerial vehicle (UAV) systems. In this design, the weighting parameters of the Neural Network (NN) are updated by using the Extended Kalman Filter (EKF). Online adaptation of these weighting parameters helps to detect abrupt, intermittent, and incipient faults accurately. We apply the proposed fault detection system to a nonlinear dynamic model of the WVU YF-22 unmanned aircraft for its evaluation. The simulation results show that the new method has better performance in comparison with conventional recurrent neural network-based fault detection strategies.

A Machine Learning Approach for Fault Detection in Vehicular Cyber-Physical Systems

A network of vehicular cyber-physical systems (VCPSs) can use wireless communications to interact with each other and the surrounding environment to improve transportation safety, mobility, and sustainability. However, cloud-oriented architectures are vulnerable to cyber attacks, which may endanger passenger and pedestrian safety and privacy, and cause severe property damage. For instance, a hacker can use message falsification attack to affect functionality of a particular application in a platoon of VCPSs. In this paper, a neural network-based fault detection technique is applied to detect and track fault data injection attacks on the cooperative adaptive cruise control layer of a platoon of connected vehicles in real time. A decision support system was developed to reduce the probability and severity of any consequent accident. A case study with its design specifications is demonstrated in detail. The simulation results show that the proposed method can improve system reliability, robustness, and safety.

A Novel Sensor Fault Detection in an Unmanned Quadrotor Based on Adaptive Neural Observer

Prompt detection and isolation of faults and failures in flight control systems are crucial to avoid negative impacts on human and environmental systems, and to the system itself. In this study, a new scheme based on a nonlinear dynamic model is designed for sensor fault detection and isolation in an unmanned aerial vehicle (UAV) system. In the proposed design, a neural network is used as an observer for faults in the UAV sensors. The weighting parameters of the neural network are updated by the Extended Kalman Filter (EKF). The designed fault detection (FD) system is applied to an unmanned quadrotor model, and the simulation results show that the proposed design is capable of the prompt detection of sensor faults.

Security Challenges of Networked Control Systems

Networked control systems (NCSs) are created by the integration of advanced communication networks, control systems, and computation techniques. This integration enhances efficiency and reliability at the expense of increased complexity and reduced security . For example, the reliance of NCSs on communication networks exposes these systems to attack vectors targeting generic networks. This chapter is an overview of pervasive NCSs’ applications, recent attacks on NCSs, and attack detection techniques. A mathematical framework for an NCS under common types of attack is presented, i.e., denial of service (DoS), false data injection (FDI), and time delay switch (TDS) attacks. Thereafter, the framework is used to developed an algorithm based on adaptive channel allocation and state estimation techniques to compensate for the destabilizing effects of TDS and FDI attacks simultaneously. Finally, the proposed algorithm is used in a case study to show the effect of injected attacks on different parts of an NCS and the capabilities of the detection algorithms. Simulation results show the algorithm can accurately detect attacks and can overcome the attack effects by adapting the communication channels.

Detection of false data injection attack on load frequency control in distributed power systems

The False Data Injection (FDI) attack on Load Frequency Control (LFC) caused by the adversary can destabilize the power system. This could cause potential economic and life damages. Therefore, the real-time detection of FDI attacks is necessary and essential to compensate negative effects of such attacks. This paper presents a neural network-based detection (NND) approach to estimate and detect the FDI attacks injected to sensing loop (SL) of the system. A two-area distributed power system is considered as our case study to demonstrate the effectiveness of NND strategy. The simulation results clearly show that the FDI attack can be detected and estimated in real-time with sufficient accuracy.

Adaptive Neural Network Based Fault Detection Design for Unmanned Quadrotor under Faults and Cyber Attacks

The occurrence of faults and failures in flight control systems of unmanned aerial vehicles (UAVs) can destabilize the system which could cause potential economic and life losses. Therefore, its necessary to detect faults and attacks in real time and modify the control system based on the occurred fault. In this paper, a neural network-based fault detection (NNFD) approach is introduced to detect and estimate the faults and false data injection (FDI) attacks on the sensor systems of a quadrotor in real time. An unmanned quadrotor is selected as our case study to demonstrate the effectiveness of our proposed NFDD strategy. The simulation results show that the applied NNFD method can detect the faults and FDI attacks on an unmanned quadrotor sensors with sufficient accuracy.