In this paper, an interval observer for continuous linear time-invariant systems with unknown input and bounded disturbances is proposed. Usually, the design of such observers is based on the cooperative property of the considered system dynamics, which is hard to satisfy in many cases. To overcome this issue, in a recent work, it has been shown that under some assumptions, the cooperativity of the studied system can be ensured by means of a time-varying change of coordinates. However, a constructive method for the design of the observer gain, making the observation error dynamics simultaneously positive and stable, is still missing and remains an open problem, especially in the context of unknown input. In this paper, a constructive approach is provided to obtain not only the observer gain but also a new gain that cancels the unknown input of the system and reduces the effect of the bounded disturbances. The proposed approach also allows for estimating the bounds of the unknown input. The efficiency of the proposed methodology is shown through numerical simulations and is compared with simulations resulting from a new method proposed in a recent work.

Kazantzis–Kravaris/Luenberger (KKL) or Nonlinear Luenberger observers are powerful theoretical tools for estimating the state of a large class of nonlinear systems. Such observers rely on the existence of an injective transformation towards new coordinates in which the original system dynamics is transformed into an exponentially stable linear filter of the output, which facilitates the determination of an estimate. The injectivity of the transformation enables to use its left inverse in order to recover the original system state estimate in the original coordinates.The present paper proposes a KKL-based interval observer, providing lower and upper bounds of the state of a discrete-time, time-varying nonlinear system affected by additive bounded disturbances and measurement noise, under the assumptions of the invertibility of the dynamics and the uniform Lipschitz continuity of both the dynamics inverse and the output map sequences. The design of this observer mainly relies on the assumption of uniform Lipschitz backward distinguishability enabling to bound the state, regarding the uncertainties, using an interval observer for the linear filter in the new coordinates. Back to the original coordinates, the inclusion of the system state into the estimated interval is then guaranteed after a finite time.

This paper presents a continuous-discrete adaptive observer for jointly estimating the states and unknown parameters of nonlinear systems with sampled output measurements. The observer combines a state estimator, a parameter estimator, and a reset-based inter-sample output predictor. By analyzing the interconnected error dynamics through an Input-to-State Stability (ISS) framework and the small-gain theorem, we establish sufficient stability conditions and derive an explicit computable bound on the Maximum Allowable Sampling Period (MASP). The proposed methodology is applied to a vehicle longitudinal dynamics model with uncertain mass, showing accurate state and parameter estimation under sampled-data measurements and bounded disturbances.

This study integrates the adaptive neural network technique, state and disturbance observer, and nonsingular terminal sliding mode control framework to achieve rapid trajectory tracking for a quadrotor under state uncertainty, dynamic disturbances, and unknown dynamics. The control methodology begins by modeling the quadrotor's dynamics, including inherent uncertainties and disturbances. A state observer is designed to estimate unknown states while rejecting uncertainties and disturbances, with exponential convergence of observer errors demonstrated via Lyapunov theory. The tracking error between the observed quadrotor's states and the desired trajectory is subsequently defined, and a nonsingular terminal sliding mode surface is employed to stabilize this error. Using Lyapunov-based analysis, it is proven that the sliding surfaces converge to zero at an exponential rate. Furthermore, an adaptive neural network strategy is developed to approximate unknown components, including state variables, uncertainties, and disturbances. Simulation and experimental results using a quadrotor with realistic parameters validate the proposed control strategy, confirming its robustness and effectiveness.

This study proposed a dynamic event-triggered (ET) observer for nonlinear systems. The observer structure is characterised by the incorporation of a closed-loop inter-sample output predictor structure in the event-triggered mechanism (ETM). The design features two main characteristics. The first one is the asymptotic convergence of the observer with respect to the designed ETM. The second one is the Zeno-free behaviour, both of them have been proven using Lyapunov theory. Two application examples are given to illustrate the effectiveness of the proposed observer.

This paper presents a novel observer-based synchronization framework designed for heterogeneous multi-agent systems operating under realistic conditions, such as discrete and noisy measurements, intermittent communication, and agent heterogeneity. The framework incorporates a hierarchical structure: a Filtered High-Gain Observer to reconstruct the leader's state in the presence of noisy and discrete measurements, a Distributed Observer which enables each follower to independently estimate the state of the leader for synchronization purposes, and a Adaptive Local Observer for each follower that simultaneously estimates its own states and unknown parameters. The local observer incorporates a closed-loop output predictor to address challenges arising from discrete measurements and ensure continuous estimation. Stability and robustness are shown using a Lyapunov-based approach, and simulation results validate the effectiveness of the proposed method to maintain precise coordination between vehicles. This approach improves performance, making it suitable for real-time vehicle platooning applications.

This study presents a novel event-triggered joint adaptive high-gain observer design for delayed output-sampled nonlinear systems with output injection. These systems are characterized by the presence of unknown parameters that influence both the state and output equations. The major difficulty in designing the observer lies in the interplay between event-triggered mechanism, output injection, and time-varying delays. Additionally, the non-affine nature of the parameter's entry into the system states equation further complicates the design. To address these challenges, a new adaptive law for unknown parameter estimation is developed under delay measurement. A novel non-Zeno dynamic event-triggered mechanism coupled with a closed-loop output predictor is proposed. The resulting observer exhibits two main features: the first one provides an input-to-state stable property, and the second one is the establishment of a theoretical condition for the inter-event time of the proposed dynamic event-triggered mechanism. The effectiveness of the designed observer is demonstrated through numerical simulations and performance comparisons with previous works.

This paper presents a novel sampled-data observer for robotic manipulators subject to internal and external disturbances. The proposed approach is a novel continuous-discrete extended high-gain observer featuring: (i) simultaneous estimation of both system states and disturbances; (ii) explicit consideration of sampled-data measurements, including potential differences in sampling frequencies between the controller and the measurement device, which provides robustness to variations in sampling; (iii) a new type of time-varying gain that significantly reduces the effect of measurement noise on the estimated variables; and (iv) a detailed theoretical analysis that guides parameter tuning to enhance the observer’s estimation accuracy. The proposed observer has been validated through real-time experiments on ERRIC industrial manipulator. The obtained results show clearly the benefits of the proposed observer design and its effectiveness as well as robustness, compared to other solutions for the literature.

The lateral state information of vehicles is pivotal for their lateral control and safety monitoring of motion states. In light of the high cost associated with a vehicle's lateral speed sensor, this study focuses on estimating the lateral velocity of autonomous vehicles using the vehicle's longitudinal speed and yaw rate. The main difficulty is that the state estimation is affected by sampled and delayed sensor data measurements as well as unknown modeling uncertainties. To overcome these challenges, this article proposes a novel sampled data neural network observer. The proposed observer consists of two components: a continuous state observer and a compensating injector. The compensating injector is devised to offset the information loss during sampling by digital sensors. Additionally, a radial basis function neural network is employed to approximate the unknown dynamics and modeling uncertainties in the vehicle systems. The weights of the neural network are updated using a newly designed weight update law. Furthermore, to tackle the challenge of both sampling and delay measurement, an additional integrator is incorporated in the designed compensating injector. This additional integrator aims to mitigate the impact of sensor data delays on the estimation process. Finally, the stability of the proposed method is demonstrated using the Lyapunov technique. The effectiveness of the proposed sampled data neural network observer is verified through simulation and experiment performed on a full driving automation vehicle experimental platform.

This paper presents a model predictive controller for autonomous vehicles based on ultra-local models and a Nonlinear Extended State Observer (NESO) that simultaneously address path tracking performance and roll stability. The controller consists of several key components: first, ultra-local models are developed to represent the vehicle lateral error dynamics and roll dynamics. Next, a nonlinear extended state observer is designed to estimate integrated disturbances and unmodeled states within the ultra-local models. The parameters of the observer are optimized using the Butterfly Optimization Algorithm (BOA) in a simulation environment to enhance estimation accuracy. Finally, a model predictive control (MPC) scheme is implemented to minimize both lateral tracking error and roll angle, effectively addressing both path tracking performance and roll stability. The effectiveness of the proposed controller is validated through comprehensive comparative simulations conducted on a MATLAB/Simulink and Carsim co-simulation platform.

This paper investigates the characteristics and performance of deep learning-based visual odometry for the localization of mobile robots in 2D indoor small-scale environments. Our study begins by developing a highly accurate multi-sensor fusion localization method that integrates data from various sensors, including cameras, inertial measurement units (IMUs), Indoor Positioning System (Marvelmind) and wheel encoders. Using this method, we created a comprehensive dataset that captures the robot’s movements in controlled indoor settings. Then, We conducted an extensive comparative study of several deep learning-based visual odometry methods by evaluating their strengths and weaknesses using public datasets. From this comparison, we identified the Deep Patch Visual Odometry (DPVO) method as the most effective approach. Afterwards, we made several enhancements to the DPVO method to further improve its localization capabilities. Subsequently, we applied the improved DPVO method to our newly created dataset, which allowed us to perform rigorous testing and validation in realistic indoor scenarios. The results were compared with localization data obtained from other modalities.

This study presents a novel approach for esti-mating lateral velocity, an important parameter for vehicle stability characterization. Aiming to resolve the problems of poor estimation accuracy caused by the insufficient modeling of traditional model-based methods and issues with sampled and delayed measurements, a sampled delay data neural network method for lateral velocity estimation is designed. Our approach incorporates a compensating injector to fill information gaps between samples, an extended compensation dynamic to reduce delays' impact, and a radial basis function neural network to mimic vehicle motions. Continuous weight updates ensure adaptability, and stability is demonstrated using the Lyapunov methodology. Experimental results confirm the effectiveness of our approach, providing promising insights to enhance lateral velocity estimation and improve control and stability in autonomous vehicle systems.

Accurately estimating the lateral velocity of automatic ground vehicles is a complex task, especially when faced with sensor-sampled measurements and unfamiliar mathematical models. In order to overcome these difficulties, the study presented here proposes a novel approach that makes use of a sampled-data neural network observer. In order to fill in the information gap between successive samples, a compensating injector is introduced to the continuous state observer on which the observer is based. In order to replicate unknown dynamic vehicle systems, a radial basis function neural network is also implemented. A special weight update mechanism is used to update the weights continually. The Lyapunov methodology is used to demonstrate the stability of the suggested method. Experimental findings validate the effectiveness of the sampled-data neural network observer, providing promising insights for improving lateral velocity estimation and enhancing the control and stability of autonomous vehicle systems.

This paper proposes a novel robust tracking control scheme for discrete time linear uncertain Multiple -Input Multiple -Output (MIMO) systems subject to time -varying delay on the states. The considered system is affected by unknown but norm bounded uncertainties on parameters as well as matched disturbances on the states. The designed controller is based upon a proposed novel integral sliding surface and a new switching type of reaching law. Sufficient conditions based on Linear Matrix Inequalities (LMIs) and a suitable Lyapunov- Krasovskii Functional (LKF) are derived in order to guarantee the asymptotic stability of such system. The proposed controller ensures a good tracking performance despite the presence of the time varying delay and the matched/unmatched disturbances. Moreover and thanks to the proposed integral surface, the time reaching phase is eliminated and the chattering phenomenon is significantly reduced. The proposed controller is applied on an Autonomous Underwater Vehicle (AUV) to follow a prescribed desired trajectory. The simulation results illustrate the effectiveness of such controller.

Three-dimensional (3D) real-time object detection and tracking is an important task in the case of autonomous vehicles and road and railway smart mobility, in order to allow them to analyze their environment for navigation and obstacle avoidance purposes. In this paper, we improve the efficiency of 3D monocular object detection by using dataset combination and knowledge distillation, and by creating a lightweight model. Firstly, we combine real and synthetic datasets to increase the diversity and richness of the training data. Then, we use knowledge distillation to transfer the knowledge from a large, pre-trained model to a smaller, lightweight model. Finally, we create a lightweight model by selecting the combinations of width, depth & resolution in order to reach a target complexity and computation time. Our experiments showed that using each method improves either the accuracy or the efficiency of our model with no significant drawbacks. Using all these approaches is especially useful for resource-constrained environments, such as self-driving cars and railway systems.

This paper proposes a novel Neural Network Adaptive Observer (NNAO) for Nonlinear Systems with Partially and Completely Unknown Dynamics (NSPCUD), subject to variable sampled and delayed output. The method involves designing a neural network observer for partially unknown nonlinear systems with sampled and delayed outputs, using a radial basis function (RBF) neural network to approximate the system’s unknown part. A new weight update algorithm is proposed, along with a closed-loop output predictor for coping with variable samples, and a closed-loop integral compensation to handle variable delay. This approach is then extended to cover completely unknown systems as well. Numerical simulations and comparisons between the proposed method and previous methods on autonomous ground vehicle models were conducted to verify the effectiveness of the proposed NNAO.

In this paper, a new High-Gain Interval Observer (HGIO) structure and its filtered version, named Filtered High-Gain Interval Observer (FHGIO), are proposed for a class of Linear Parameter Varying (LPV) systems subject to additive disturbances and measurement noise. Those uncertainties are assumed to be unknown but bounded with known values. The HGIO is based on a high-gain observer structure from which an interval formulation is deduced taking into account the uncertainties bounds. Then, the proposed HGIO is extended to incorporate a filter for the output estimation error, leading to the FHGIO design whose goal is to reduce the measurement noise amplification. Usually, the design of such interval observers is based on monotone systems theory which is hard to satisfy in many cases. In this paper, suitable changes of coordinates are used to overcome this limitation. Moreover, a sufficient condition for the non-divergence of the radius dynamics and a procedure to design the observers gains ensuring the stability are given for each observer. The efficiency of the proposed observers is illustrated through a simulation on a numerical example.

This paper introduces a novel adaptive neural network-based sliding mode controller for trajectory tracking of a quadrotor UAV under unknown parameters such as inertia and mass. Unknown external disturbances and nonlinear aerodynamic forces are also considered. Due to its complex dynamics, the quadrotors need a dual controller for outer and inner loop control. For the commercial UAV, the inner loop controller, which ensures the attitude control, is usually assumed to be well developed and unmodifiable. Under this assumption, this paper focuses on the position control of a quadrotor UAV. A new position controller will be then designed and used to compute the appropriate input commands for the attitude controller to achieve the trajectory tracking task. The proposed method combines both Back-Propagation Neural Network (BPNN) scheme and sliding mode control to eliminate the effects of exogenous disturbances and model uncertainties. To illustrate the efficiency of the proposed controller, a comparative analysis is performed with different methods through experimental results.

In this article, a novel structure of a filtered high-gain observer (FHGO) is proposed for a class of nonlinear systems subject to sampled and delayed measurements, bounded disturbances, and noise measurements effects. Using an using an linear matrix inequality (LMI) design approach, the novel structure of the continuous-discrete FHGO allows a larger bound of the maximum allowable value of time delay. To prove the effectiveness of the proposed observer, a comparison with Kalman-like observer and standard high gain observer is provided and the corresponding experimental results are presented for a quadrotor UAV.