Keywords: Autonomous vehicles, Transportation systems, Automobile
Abstract: This paper presents a design method for the cruise control of connected automated vehicles. The proposed control system contains layers with different functionalities, which are interconnected in a hierarchical structure. The reason for using layers is to separate the functionalities which require different computation efforts. The layer on the high level computes a reference speed signal for the vehicle, which considers the economy and time performance requirements. It requires the solution of a complex optimization problem, which is performed in a cloud with low frequency. There are two further layers on the level of the vehicle, such as the signal analysis and the local control layers. The purpose of the signal analysis layer is to examine the feasibility of the reference speed regarding vehicle safety performances. This layer is designed to require low computation effort and on-board sensor measurements, thus analysis can be performed with high frequency. The proposed hierarchical cruise control is able to handle the priorities between the various performance requirements, thus safe motion of the vehicle is guaranteed under every circumstances.

Keywords: Automobile, Fault tolerant control / fault recovery
Abstract: This paper proposes a lateral Advance Driver Assistance System (ADAS) steering controller that uses the detection of driver errors to reconfigure a Linear Parameter Varying (LPV) controller acting on the combined driver-vehicle lateral steering system. The detection of driver errors is proposed to be formulated as a Fault Detection problem using the Parity Space approach, and the computed residual signal then schedules the LPV/Hinf controller. The overall goal is to compute an ADAS controller that helps in stabilizing the vehicle when driver errors are detected while otherwise minimizing the level of intrusiveness. With this goal in mind, the proposed method was tested in simulation using a full dynamical model of a Renault Megane car and driver models, for simulation of the human steering action, during a critical scenario.

Keywords: Fault detection and isolation, Design for reliability and safety, Communication networks
Abstract: Currently the operation of modern industrial processes depends on many different components (systems and plant devices) all of which are tied together through communication systems. In large scale systems, due to the difficulty of monitoring the communication infrastructure, the latter is vulnerable to external physical intervention, which could have as a consequence the loss of communication between process components. In some cases, to avoid loss of communication, redundancies or alternative paths are implemented, but this is not necessarily the most efficient solution. In this paper is proposed a fault diagnosis approach based on structural analysis that makes effective use of its infrastructure in communication-loss condition. This is accomplished by considering the communication structure in the proposed approach in order to apply optimization methods guaranteeing optimal use of the available communication structure. For this, diagnosers choose the best communication paths or, if necessary, change their design. The proposed approach is applied to a power distribution system achieving diagnosis goals in communication-loss condition.

Keywords: Model-based methods, Cybersecurity, Fault tolerant control / fault recovery
Abstract: The complexity of modern control systems requires mitigating the effects of numerous anomalies, such as attacks, faults and network-induced disturbances. In this perspective, the synthesis problem of a dissipation block for anomaly-tolerant control is addressed. The dissipation technique consists in interconnecting a block to a system in order to make their interconnection dissipative. The underlying idea is that, if the resulting dissipative characterization is suitably chosen, the considered anomalies can be mitigated. In this paper, it is revealed how the block, here an LTI system, can be synthesized by deriving sufficient LMI-based conditions. A numerical example is provided to illustrate this result which is applied to design a control reconfiguration, providing a defense mechanism that ensures resilience with respect to a data-injection cyber attack.

Keywords: Fault tolerant control / fault recovery, Data driven methods, Autonomous vehicles
Abstract: This paper develops a transfer-learning approach in a reinforcement learning framework for fault-tolerant control of a class of dynamical systems. Our approach combines policy-reuse and consolidation algorithms to achieve learning speed improvement when new faults occur in the system. Our policy-reuse strategy finds a policy parameter initialization that leverages previous knowledge in the form of learned policies for past fault occurrences. We present the conditions under which our approach becomes effective and empirically demonstrate our approach on a test-bed of a 6-tank fuel transfer system of an aircraft.

Keywords: Fault tolerant control / fault recovery, Data driven methods, Model-based methods
Abstract: Fault-tolerant control (FTC) focuses on developing algorithms to accommodate the impact of system faults while allowing the system to continuously operate in a degraded manner. Additionally, data-driven methods like reinforcement learning (RL) have shown excellent performance for complex continuous control tasks. However, faults affect the system dynamics which disrupt optimal policy guarantees as the Markov Property cannot be satisfied. In this work, we propose a scheme based on a combination of parameter estimation, RL, and model-based control to handle faults in a continuous control environment. We empirically demonstrate our approach on a complex octocopter trajectory-tracking task subject to single and multi-motor faults. We show improved performance compared with nominal hierarchical PID control for large magnitude faults. Lastly, we demonstrate our approach’s robustness against noisy parameter estimation.

Keywords: Maintenance policies, Design for reliability and safety, Fault tolerant control / fault recovery
Abstract: By applying the initial state design using a state-dependent switching L2 gain to the safe maintenance procedure presented before, we improve the safety of transitions between operating states to bring the procedure to completion as a support technology for safe and partially-performed preventive maintenance of linear time-invariant control systems.

Keywords: Data driven methods, Autonomous vehicles, Design for reliability and safety
Abstract: Pedestrian collision avoidance is a relevant safety aspect for autonomous driving systems operating in urban scenarios. This paper presents a Reinforcement Learning approach to endow the resulting agent with the following two competing capabilities: managing unexpected pedestrian crossings and tracking a specific trajectory. In particular, we use the Deep Deterministic Policy Gradient, a model-free off-policy algorithm for learning continuous actions. The effectiveness of the proposed Reinforcement Learning system and the associated training approach is demonstrated by means of numerical simulations.

Keywords: Design for reliability and safety, Networked control system, Autonomous vehicles
Abstract: Autonomous systems strive to obtain salient fea-tures that include computer intelligence for obtaining situationawareness, decision support to a human navigator, or for facili-tating autonomous decision-making in unmanned vehicles. Thispaper considers the case of autonomous marine surface vehicles,where high-quality decision support will be instrumental forobtaining a periodically unattended bridge and for approvalof unmanned bridge operation with fallback through remoteoperation. The proposed design focuses on a sovereign-basedarchitecture that facilitates safety, resilience and cyber-security.We address central elements of risk in the development andapproval of autonomous systems; we analyze the challengesassociated with testing, commissioning and maintenance ofa highly complex cyber-physical system, and describe designprinciples for the sovereign agents architecture.

Keywords: Cybersecurity, Fault-forecasting methods, Data driven methods
Abstract: Defense against cyberattacks is an emerging topic related to fault-tolerant control. In order to avoid difficult mathematical modeling, model-free control (MFC) is suggested as an alternative to classical control. For illustration purpose a Load Frequency Control of multi-areas power network is considered. In the simulations, load altering attacks and Denial of Service (DoS) in the communication network are applied to the system. Our aim is to compare the impact of cyberattacks on control loops closed via respectively a classical controller in such situations and a model-free one. Computer experiments show impressive results with MFC.

Keywords: Cybersecurity, Autonomous vehicles
Abstract: Nowadays, many vehicles are connected to the Internet for providing enhanced or additional functionalities to drivers. However, the existing vehicle communication protocols are not designed for cybersecurity. As a result, Autonomous Vehicles (AVs) are exposed to possible hacker attacks which could jeopardize the safety of passengers. In this paper, a novel Informative Model Predictive Scheme (I-MPS) against Replay Attacks (RAs) and Denial of Service (DoS) attacks for vehicles equipped with control systems for autonomous vehicles is proposed. The effectiveness of the scheme is shown by numerical simulations in a typical scenario of an autonomous vehicle during an overtaking and is compared against a typical architecture not implementing the I-MPS.

Keywords: Cybersecurity, Automobile, Design for reliability and safety
Abstract: The paper is focused on the identification of potential local attack threats in embedded systems used in automotive applications. One of the objectives of the paper is to show that cyber-attacks may end with taking control of a module in an electric vehicle. On the other hand, the article proposes the low-cost protection method that can be used for increasing the level of CAN bus transmission security

Keywords: Fault detection and isolation, Model-based methods
Abstract: This paper presents a new approach to the fault detection for lumped-parameter LTI systems subject to additive and multiplicative actuator, sensor and process faults. Additionally, disturbances with finite energy on a moving horizon are taken into account. A residual generator using only simple moving horizon integrals of known signals is derived. For the residual generator design a systematic method is proposed, which is based on a flatness-based trajectory planning problem. Thereby, free parameters are determined to achieve a maximized sensitivity of the residual signal to the fault in the presence of disturbances. The results are demonstrated for a numerical example and compared to results of an observer-based fault detection approach taken from the literature. The corresponding simulation results validate the increased fault detection performance gained by the new approach.

Keywords: Fault detection and isolation, Health monitoring, Model-based methods
Abstract: When attempting to meticulously conduct fault detection and isolation, inferential sensing approaches can prove to be valuable tools, as they constitute cost-effective and reliable alternatives to expensive and often impractical measuring devices. In this work, we combine advanced model-based sensor selection and inferential sensing techniques to yield accurate fault detection outcomes, in the presence of system noise and uncertainty. In this regard, the most informative sensor set is chosen, by comparing all possible sensor combinations, based on various optimality criteria that extract the maximum knowledge from the system in functions of the Fisher Information Matrix. Subsequently, optimal inferential sensors are derived, by means of genetic and mathematical programming that incorporate symbolic regression and optimization techniques. For built-in test deployment, k-Nearest Neighbors (k-NN) classification is employed to assess the accuracy of each sensor network, as well as the performance enhancement due to the inclusion of inferential sensor(s). The proposed methods are applied in steady state and dynamic models of a cross-flow plate-fin heat exchanger system for various levels of measurement noise and uncertainty. The results illustrate that the augmented system of composite sensors (i.e., inferential and hardware) provide more accurate information on system fault(s), while reducing the evidence of uncertainty and system noise.

Keywords: Fault detection and isolation, Model-based methods
Abstract: This paper presents the performance of zonotopic fault detection (FD) for additive and multiplicative fault using direct test and inverse test. Zonotopic set-based approaches use a zonotope to describe the uncertain state, parameter and noise based on an unknown but bounded description. These FD test methods aim at checking the consistency between the measured and estimated behaviour obtained from estimator in the parameter or output space. When an inconsistency is detected between these two, a fault can be indicated. At last, a motor model will be used to compare the performance of direct test and inverse test for additive and multiplicative faults.

Keywords: Fault detection and isolation, Model-based methods, Power plants / energy transport
Abstract: This paper describes a diagnosis scheme based on a distributed adaptive observer identifying a shunt fault of a long transmission line. A conductance deterioration in a specific point is assumed. Considering that the hyperbolic partial differential equations govern the electric transmission line, fault location and estimation are obtained by means of an adaptive distributed observer with only available measurements at the boundary. The key ingredient in the approach is the separation of the fault magnitude in the model from its position into two consecutive tasks. This is achieved by means of an adequately selected canonical transformation of original variables and an adaptive observer design for the transformed system. The effectiveness of the approach is illustrated by a numerical simulation in a long electrical transmission line of 300km.

Keywords: Health monitoring, Communication networks, Transportation systems
Abstract: The massive deployment of electronic components in various transportation systems has increased the complexity of their embedded networks. Power line communication is a good candidate for reducing cable bundles in these systems. Even if the number of cables is reduced, these cables tend to encounter faults that can lead to total system failure. Therefore, it is critical to design and implement an effective soft fault detection system. In this paper, a Y-shaped network consisting of one source and two receivers is studied. A soft fault, which can be caused by a localized degradation of the cable quality, is represented by a series resistance. The transmission transfer function measured at each receiver is monitored. A new health indicator inspired by the classical cable chain matrix model is proposed. This indicator is calculated at each receiver. A residual based on the calculated indicator is then used along with a preconstructed topology dependent signature matrix to detect a soft fault and to locate the affected branch. Real data extracted from a test bench are used to validate our approach. The results confirm the ability to use Power Line communication systems for diagnostic purposes in addition to their initial implementation purpose.

Keywords: Fault tolerant control / fault recovery, Autonomous vehicles, Mechatronic and robotics
Abstract: Fixed-wing UAVs show a high interest and high challenges in control, fault diagnostic, and fault-tolerant control. Accurate models of highly coupled nonlinear systems as the Fixed-Wing UAV are complex and time consuming. A need to find a control architecture independent of the model, suitable for nonlinear systems and can reject disturbances is highly desirable. One example on such a control approach is the concept of Model Free Controllers (MFC). This paper compares MFC architecture to the standard PI controller applied to a Fixed-Wing UAV in fault free case, trajectory tracking in the presence of wind disturbance, and fault tolerance in case of loss of effectiveness of actuators. The work consists of implementing the controllers in MATLAB/Simulink on a fixed-wing UAV model.

Keywords: Fault detection and isolation, Fault tolerant control / fault recovery, Aeronautics / aerospace
Abstract: In the literature several control schemes exist which utilise knowledge of an actuator’s saturation limits to ensure feasible control signals and prevent the associated performance loss caused by saturating actuators. Typically these works make an assumption that the saturation limits are known. To increase the applicability of these schemes to scenarios where the saturation limits may vary (due to faults/failures or changes in the operating conditions), this paper proposes a fault detection and isolation scheme which is able to estimate both an actuator’s health and its saturation limits. The proposed scheme is tested within an optimised control allocation and sliding mode control framework for robust fault tolerant control of a Prandtl wing aircraft.

Keywords: Fault tolerant control / fault recovery, Aeronautics / aerospace
Abstract: This paper introduces a fault tolerant control scheme based on an LPV Sliding Mode Control approach with Control Allocation, to take advantage of the accessible rotors in an octoplane. Modelling of the octoplane and a synthesis of the controller in the presence of uncertainty as well as faults/failures, is introduced in the paper. The FTC scheme is based on an on-line control allocation approach which utilises the effectiveness level of the rotors to redistribute the control signals to the healthy vertical rotors. Simulation results show good tracking performance and no visible degradation in the presence of uncertainties and faults/failures.

Keywords: Aeronautics / aerospace, Fault tolerant control / fault recovery
Abstract: This paper details the development of a high redundancy Unmanned Aerial Vehicle (UAV) platform with a sliding mode fault tolerant control system. The high redundancy UAV platform is of a Tilt Wing design, utilising four wings, each with a propeller. Conventional fixed wing aircraft control surfaces are also implemented i.e. ailerons, elevators and rudders. The benefits of such a platform is it combines the efficient forward flight of fixed wing aircraft, with the agility and Vertical Take Off and Landing (VTOL) capability of a conventional multirotor aircraft. Results show good forward flight performance despite the presence of actuator faults/failures.

Keywords: Risk analysis, Design for reliability and safety, Fault tolerant control / fault recovery
Abstract: Unmanned Aerial Vehicles (UAVs) undergo fast-growing integration into civilian airspace. Recently, the European Aviation Safety Agency (EASA) published regulations forming top-level safety targets for certified categories for UAVs. In this paper, we present a qualitative criticality analysis for a simplex electromechanical actuator (EMA) to evaluate the reliability of its components. Consequently, several hardware redundancies have been iterated to qualify the actuator architecture for the safety requirements and technical constraints. Finally, we verify the possible operating modes for the EMA using stochastic Markov models. Markov models have been also utilized for exploring health monitoring strategies for fault-tolerant features and their operating modes.