This paper investigates robust cooperative control strategies for multi-robotic car systems operating under sensor and actuator faults. In autonomous driving environments, the degradation or failure of sensors and actuators significantly affects the performance of the system, posing risks to formation control, velocity tracking, and safety. To address these challenges, we propose a robust neural control framework that integrates a dynamic adjustment neural network (DANN) with fault-tolerant design. This architecture enables each robotic car to adaptively learn the system dynamics and adjust control signals in real time, even in the presence of component faults. A fault detection and isolation (FDI) mechanism is incorporated to identify malfunctioning elements, allowing the control system to dynamically compensate and maintain coordinated behavior. Lyapunov-based analysis is employed to guarantee stability and convergence of the system. In addition to theoretical development, a detailed simulation example involving a team of robotic cars under various sensor and actuator fault scenarios is presented to demonstrate the effectiveness and robustness of the proposed control strategy. The results confirm reliable tracking performance, strong resilience, and improved formation stability under realistic fault conditions.

In this paper, we addressed the issue of maintaining a desirable level of performance in the presence of actuator faults and deception attacks in nonlinear multi-agent systems (MASs). These problems are critical for the stability and coordination of MASs that are increasingly used in robotics, autonomous vehicles, and industrial automation. Our aim was to design control strategies that, in the presence of these challenges, guarantee practical finite-time stability and robust tracking performance. To accomplish this, a distributed adaptive fuzzy control scheme based on backstepping was developed. Fuzzy logic systems were utilized to capture the complex system's unknown nonlinearities, while adaptive laws were designed to estimate and mitigate actuator gain and bias faults. A Nussbaum-type function was introduced to address unknown control directions resulting from deception attacks. Stability was verified by the Lyapunov theory. The suggested approach ensured that it was a finite-time stable method, and every signal in the closed loop was found to be semi-globally uniformly eventually bounded. Our control strategy, compared to published approaches, improved convergence time by approximately 38% and tracking accuracy of approximately 35% under the same conditions of simultaneous actuator faults and deception attacks.

Training deep neural networks remains difficult due to vanishing gradients, non-convex loss surfaces, and hyperparameter sensitivity. These obstacles are compounded by quantum machine learning, where barren plateaus, circuit depth, and hardware noise restrict the applicability of gradient-based approaches. To overcome these drawbacks, this study presents adaptive Grover-driven parallel quantum optimization (AG-PQO), a hybrid, gradient-free scheme that leverages Grover's quadratic search speedup, along with adaptive loss-aware discretization and fidelity-based regularization. In contrast to more classical optimizers, such as Adam or evolutionary strategies (ES), which are either sensitive to the adequacy of the gradient update or exhibit poor scaling behavior, AG-PQO optimizes by performing Grover-accelerated candidate exploration across layers and reuses high-quality solutions in quantum memory caching. Testing indicates that AG-PQO yields higher accuracy, 2%–3% above Adam and ES, and faster convergence with less end-value loss than Adam, ES, and quantum feedforward-backpropagation (QFB). It is worth noting that AG-PQO remains stable at the simulated noise level of NISQ and has the potential to scale to near-term quantum processors.