This paper investigates observer-based event-triggered sliding mode control of delayed semi-Markov jump discrete singular systems under randomly occurring false data injection attacks and network-induced communication restrictions. Error dynamics and observer systems are modeled as semi-Markov jump systems with time-varying time delay using an improved time delay analysis technique. The stochastic admissibility of the augmented system is analyzed using the semi-Markov kernel method, and adequacy criteria that facilitate numerical detection are derived from new Lyapunov functions. A mode-independent sliding surface is designed and its reachability is ensured. Finally, the effectiveness of the proposed scheme is illustrated by simulation examples.

In this article, we propose a novel strategy leveraging multiresolution topological maps to guide ground robots in exploring complex 3-D environments. Our approach extracts traversable areas and local frontiers from detailed elevation maps and integrates them into a global multiresolution topological map with global frontiers and viewpoints. GPU-accelerated computation enhances efficiency, while a viewpoint-guided adaptive sampling mechanism generates a sparse global map for rapid path-length estimation to viewpoints. Collision-free global paths to target areas are computed efficiently using the multiresolution topological map. Extensive simulations with wheeled and legged robots demonstrate superior task completion rates and exploration efficiency compared to state-of-the-art methods, with further validation in real-world experiments. The code will be made publicly available to support the robotics community.

There is a lack of rigorous stability concept which can stipulate the actual settling time of a dynamic system in existing studies. In this article, a stipulated-time stability for nonautonomous dynamic systems is proposed, which is then extended to stipulated-time boundedness for uncertain systems. By using a class of bounded time-varying functions, Lyapunov-like conditions to ensure a dynamic system to exhibit stipulated-time stability/boundedness are developed. It is interesting that previous Lyapunov-like theorems for predefined-time (PDT) stability can be unified into our framework to achieve stipulated-time stability. To validate the framework, a stipulated-time controller is first designed for a general affine system, which requires a smaller initial control signal than that of PDT control. Furthermore, a generalized design of stipulated-time distributed observer for leader-following multiagent systems is proposed.

Sequential decision-making can be formulated as a conditional generation process, with targets for alignment with human intents and versatility across various tasks. Previous return-conditioned diffusion models manifest comparable performance but rely on well-defined reward functions, which requires amounts of human efforts and faces challenges in multi-task settings. Preferences serve as an alternative but recent work rarely considers preference learning given multiple tasks. To facilitate the alignment and versatility in multi-task preference learning, we adopt multi-task preferences as a unified framework. In this work, we propose to learn preference representations aligned with preference labels, which are then used as conditions to guide the conditional generation process of diffusion models. The traditional classifier-free guidance paradigm suffers from the inconsistency between the conditions and generated trajectories. We thus introduce an auxiliary regularization objective to maximize the mutual info

This article delves into the predefined-time output-feedback leader-following consensus problem of uncertain pure-feedback nonlinear multiagent systems for the first time. To streamline subsequent design, the original systems in pure-feedback form are first transformed into canonical systems. Following this, a distributed predefined-time extended state observer (ESO) and a local predefined-time ESO are developed to reconstruct the unknown states/lumped disturbance of the transformed leader system and follower systems, respectively. Based on the estimated states and utilizing a bounded regulation function, two nonsingular and nonconservative predefined-time control laws are formulated to achieve consensus tracking. The proposed method showcases the following advantages: 1) the actual convergence time rather than the upper bound of the convergence time (UBCT) of the tracking errors can be explicitly specified a priori regardless of the initial conditions in a bounded region, optimizing control energy usage and 2) the system overshoot could be effectively reduced by selecting appropriate parameters for the regulation function. Finally, numerical examples are conducted to verify the obtained results.

The fractional-order (FO) nonlinear differential system with state-dependent (SD) delayed impulses (DI) is considered in this brief. The considered impulses are related to the delayed state of the system and the delays are SD. A novel lemma for the monotonicity of the solution of Caputo's FO derivative equation is given. By means of linear matrix inequality (LMI) and several comparative arguments, criteria of uniform stability, uniform asymptotical stability, and Mittag-Leffler stability are obtained. Compared with other works on integer-order (IO) impulsive delayed systems with SD delays or fixed delays, how to impose constraints on parameters and impulses is explored, without imposing the boundedness on the state delays. Two examples are implemented to examine the practicality and sharpness of our theoretical analysis.

To handle the tracking problem of uncertain strict-feedback nonlinear systems with matched and mismatched composite disturbances, this article studies a predefined-time backstepping controller by resorting to a Lyapunov-based predefined-time dynamic paradigm, a regulation function, and neural networks (NNs). Moreover, an adding-absolute-value (ADV) technique is adopted in the design process to remove the control singularity. Theoretical analyses prove the boundedness of all closed-loop system signals and the predefined-time convergence of the tracking error into an arbitrarily small vicinity of the origin. The proposed controller exhibits four advantages: 1) the actual convergence time is precisely predefined by only one design parameter irrespective of the initial conditions, and the control energy is economized; 2) no unbounded terms are adopted for predefining the actual convergence time, thus avoiding numerical overflow problem under limited memory space and gaining strong noise-tolerant ability; 3) the peaking tracking error and control input magnitude can be effectively reduced by appropriately setting parameters of the regulation function; and 4) the controller is continuous and nonsingular everywhere. Finally, a practical example of a single-link manipulator is presented to validate the efficacy and superiority of our predefined-time controller.

In this article, the finite-time observer-based sliding-mode control (SMC) problem is considered for stochastic Markovian jump systems (MJSs) with a deterministic switching chain (DSC) subject to time-varying delay and packet losses (PLs). First, the stochastic MJSs with DSC are appropriately modeled and the PLs case is characterized by using some Bernoulli random variables. Then, a nonfragile finite-time bounded sliding-mode observer is designed. Our objective is to propose a finite-time observer-based SMC approach such that for the above addressed system, the finite-time boundedness in a certain time interval can be guaranteed by giving sufficient criteria via the stochastic analysis skills and average dwell time (ADT) method. Moreover, a new robust finite-time sliding-mode controller can be designed to ensure reachability of the common sliding surface in the estimation space. Finally, a numerical example is provided to illustrate our theoretical results.

Reinforcement learning (RL) requires a lot of interactions with the environment, which is usually expensive or dangerous in real-world tasks. To address this problem, offline RL considers learning policies from fixed datasets, which is promising in utilizing large-scale datasets, but still suffers from the unstable estimation for out-of-distribution data. Recent developments in RL via supervised learning methods offer an alternative to learning effective policies from suboptimal datasets while relying on oracle information from the environment. In this article, we present an offline RL algorithm that combines hindsight relabeling and supervised regression to predict actions without oracle information. We use hindsight relabeling on the original dataset and learn a command generator and command-conditional policies in a supervised manner, where the command represents the desired return or goal location according to the corresponding task. Theoretically, we illustrate that our method optimizes the lower bound of the goal-conditional RL objective. Empirically, our method achieves competitive performance in comparison with existing approaches in the sparse reward setting and favorable performance in continuous control tasks.

This article deals with the problem of mobile robot path planning in an unknown environment that contains both static and dynamic obstacles, utilizing a reinforcement learning approach. We propose an improved Dyna-${Q}$algorithm, which incorporates heuristic search strategies, simulated annealing mechanism, and reactive navigation principle into${Q}$-learning based on the Dyna architecture. A novel action-selection strategy combining$\varepsilon $-greedy policy with the cooling schedule control is presented, which, together with the heuristic reward function and heuristic actions, can tackle the exploration-exploitation dilemma and enhance the performance of global searching, convergence property, and learning efficiency for path planning. The proposed method is superior to the classical${Q}$-learning and Dyna-${Q}$algorithms in an unknown static environment, and it is successfully applied to an uncertain environment with multiple dynamic obstacles in simulations. Further, practical experiments are conducted by integrating MATLAB and robot operating system (ROS) on a physical robot platform, and the mobile robot manages to find a collision-free path, thus fulfilling autonomous navigation tasks in the real world.

Impact mitigation is crucial to the stable locomotion of legged robots, especially in high-speed dynamic locomotion. This paper presents a leg locomotion system, including the nonlinear active compliance control and the active impedance control for the steel wire transmission-based legged robot. The developed control system enables high-speed dynamic locomotion with excellent impact mitigation and leg position tracking performance, where three strategies are applied. a) The feed-forward controller is designed according to the linear motor-leg model with the information of Coulomb friction and viscous friction. b) Steel wire transmission model-based compensation guarantees ideal virtual spring compliance characteristics. c) Nonlinear active compliance control and active impedance control ensure better impact mitigation performance than linear scheme and guarantee position tracking performance. The proposed control system is verified on a real robot named SCIT Dog, and the experiment demonstrates the ideal impact mitigation ability in high-speed dynamic locomotion without any passive spring mechanism.

In this paper, we extend a famous motion planning approach, GPMP2, to multi-robot cases, yielding a novel centralized trajectory generation method for the multi-robot formation. A sparse Gaussian Process model is employed to represent the continuous-time trajectories of all robots as a limited number of states, which improves computational efficiency due to the sparsity. We add constraints to guarantee collision avoidance between individuals as well as formation maintenance, then all constraints and kinematics are formulated on a factor graph. By introducing a global planner, our proposed method can generate trajectories efficiently for a team of robots which have to get through a width-varying area by adaptive formation change. Finally, we provide the implementation of an incremental replanning algorithm to demonstrate the online operation potential of our proposed framework. The experiments in simulation and real world illustrate the feasibility, efficiency and scalability of our approach.

In this paper, a prescribed performance control (PPC) is proposed for hypersonic flight vehicles (HFVs) in the presence of uncertainties and input saturations. A simplified error transformation is implemented to relax the constraints on the tracking performance, including overshoot, convergence rate and steady-state error. The stabilization of the transformed system is sufficient to achieve the prescribed performance. Based on the transformed error, a novel auxiliary system combined with the adaptive control is constructed to analyze and then compensate for the effect of input saturations. Simulation is provided to verify the effectiveness of the proposed scheme.

This paper mainly concerns the issue of asynchronously switched control for a class of continuous-time switched linear systems with persistent dwell time (PDT). Asynchronous switching implies that when a subsystem switches the matched controller remains active for a finite period of time, such that the Lyapunov-like function increases. A semi-time-dependent (STD) multiple Lyapunov-like function with special form suitable to asynchronously switched systems is proposed, upon which a set of STD and mode-dependent stabilizing controllers is designed. The validity and advantage of the proposed results is demonstrated through a numerical example.

In this paper, the output feedback control problem is investigated for general nonlinear industrial processes. At the device layer, the nonlinear industrial processes with disturbances are modeled by utilizing Takagi-Sugeno modeling approach, and the corresponding local controllers are then designed to guarantee that the outputs for the local subsystems can track the decomposed setpoints. At the operation layer, considering the effect of radial basis function performance index and packet dropout phenomenon, a setpoint compensator is constructed to dynamically regulate the setpoints and track the given operation index. Finally, a network-based continuous stirred tank reactor system is considered to verify the validity of the proposed strategy in the simulation part.

This work focuses on stability analysis and stabilization synthesis problem for a class of Takagi-Sugeno (T-S) polynominal fuzzy-model-based time-delay systems based on the sum-of-squares (SOS) approach. Firstly, a novel stability criterion is proposed based on Lyapunov stability theorey, which is less conservative as the information of the membership functions is included in the stability conditions. Then, a new fuzzy controller with greater design flexibility to stabilize the colsed-loop system is developed. Finally, a simulation example is provided to illustrate the effectiveness of the proposed design methods.

This study is mainly concerned with fuel efficiency modeling and prediction for common automobiles based on an informative vehicle database. The historical database is processed and the mutual information index (MII) is employed to identify a set of characteristics that significantly affect fuel efficiency. Five different machine learning techniques are exploited to build fuel efficiency prediction models. Among these techniques, quantile regression, which is a natural extension of classical least square estimation, is shown to have better performance for fuel efficiency prediction compared to other adopted techniques. It is also demonstrated that with the selected attributes based on MII, the prediction performance is almost ideal when exploiting the complete dataset.

Segmentation of neural stem cells is the preliminary step to treat and cure several brain neural diseases. There exist a number of methods to accomplish this task. However, all of these methods suffer from some problems, such as high intensity variation sensitivity, human interaction and high computational complexity. In this paper we proposed a novel edge-detection-based neural stem cell image segmentation algorithm using the local complex phase characteristics. The proposed method is an illumination and contrast invariant measurement of edge significance. Our contributions are that, local weighting summation Gaussian kernel convolution and a new model for phase deviation weighting function are introduced into the proposed model to improve the local phase measurement. In experiments, we show that the proposed method is more accurate and reliable than three existing gradient-based edge detection algorithms and Kovesi's model for neural stem cell image segmentation.