Data-based controller for a soft robot actuated by ultrasonic atomization with unknown dynamics and uncertainty

Rehabilitation robots are widely recognized as vital for restoring motor function in patients with lower-limb impairments. A Modified Fractional-Order Proportional-Integral-Derivative (MFOPID) controller is proposed to improve trajectory tracking of a 2-DoF Lower Limb Rehabilitation Exoskeleton Robot (LLRER). The classical FOPID is augmented with a modified control formulation by which steady-state error is reduced and the transient response is sharpened. Controller gains and fractional orders were tuned offline using a hybrid metaheuristic Improved Elk Herd Optimization hybridized with Grey Wolf and Multi-Verse Optimization algorithms (IElk-GM) so that exploration and exploitation are balanced. Superiority over the classical FOPID was demonstrated in simulations under linear and nonlinear trajectories, with disturbances and parametric uncertainty: 0% overshoot was achieved at both hip and knee joints; settling time was reduced from 6.998 s to 0.430 s (hip) and from 7.150 s to 0.829 s (knee); ITAE was reduced from 23.39 to 2.694 (hip) and from 16.95 to 3.522 (knee); and the hip steady-state error decreased from 0.018 Rad to 0.0015 Rad, while the knee steady-state error remained within 0.011 Rad. Control torques remained bounded under linear tracking (<345 N·m at the hip; <95 N·m at the knee) and under nonlinear cosine tracking (<350 N·m at the hip; <100 N·m at the knee). These results indicate that safer, smoother, and more effective robot-assisted rehabilitation can be supported by the proposed controller.

Self-balancing robots, with their compact size, are capable of achieving high agility. Small wheel-legged self-balancing robots have demonstrated significant potential across various applications. However, expanding small self-balancing robots to larger sizes to serve as personal transport tools is a more attractive and impactful direction than further miniaturization or confinement to niche laboratory demonstrations. This paper presents the development of a small self-balancing robot, which is then scaled up to a larger version designed to carry human passengers as a self-balancing wheelchair. A unified control framework, built around a shared core of online model-updating LQR for balance and PD for steering, is applied to both robots. This core is supplemented with platform-specific modules, such as a dedicated leg controller for the wheel-legged robot, to handle distinct dynamic maneuvers. The LQR controller is implemented for balance control in both robots. Additionally, a dedicated leg controller is applied exclusively to the small wheel-legged robot to enable dynamic maneuvers, such as jumping. A series of experiments conducted with the final prototypes validate the effectiveness of the control systems and highlight the robots’ application potential.

Abstract When legged robots impact their environment executing dynamic motions, they undergo large changes in their velocities in a short amount of time. Measuring and applying feedback to these velocities is challenging, further complicated by uncertainty in the impact model and impact timing. This work proposes a general framework for adapting feedback control during impact by projecting the control objectives to a subspace that is invariant to the impact event. The resultant controller is robust to uncertainties in the impact event while maintaining maximum control authority over the impact-invariant subspace. We demonstrate the improved performance of the projection over other commonly used heuristics on a walking controller for a planar five-link-biped. The projection is also applied to jumping, box jumping, and running controllers for the compliant 3D bipedal robot, Cassie. The modification is easily applied to these various controllers and is a critical component to deploying on the physical robot. Code and video of the experiments are available at https://impact-invariant-control.github.io/ .

A study on the optimal control strategy using sliding mode controller for Mecanum-Wheeled omnidirectional mobile robot

Purpose This study aims to realize long-horizon robotic manipulation guided by implicit instructions that convey real intent through metaphors, emotional expressions and other indirect means. Design/methodology/approach First, this study proposed shared attributes to enhance the ability of visual-language models (VLMs) in reasoning explicit intentions from implicit instructions. Specifically, the VLM was fine-tuned by adding shared attributes. These attributes are derived from those with the highest similarity extracted from images and explicit instructions, bridging intrinsic cross-modal semantic mappings between implicit expressions and explicit intentions. Owing to the lack of relevant data, an implicit instruction-based data set was constructed for fine-tuning the VLMs. Then, a hierarchical learning strategy was introduced to map explicit instructions to robotic controller parameters through a planning module, a sequencing module and an interaction learning module. Findings In these experiments, the fine-tuned VLM achieved state-of-the-art performance on both this study’s constructed data set and the public VAGUE benchmark and successfully executed ten implicit-instruction-guided robotic manipulation tasks in simulation and eight in the real world. Originality/value This work integrates implicit instructions into robot manipulation. To the best of the authors’ knowledge, this if the first study to introduce long-horizon robotic manipulation guided by implicit instructions. The authors propose to fine-tune the VLM by adding shared attributes, bridging intrinsic cross-modal semantic mappings between implicit expressions and explicit intentions. This study further introduces a hierarchical learning strategy to enable efficient transformation from implicit instructions to executable operations. This approach provides new perspectives on language-conditioned robotic manipulation and has the potential to be extended to a wide range of human-centered manipulation tasks. The video can be found at here.

During repeated rehabilitation training, the friction between the omnidirectional wheel and the desktop can lead to control response delays, causing deviations in the training trajectory. To enhance the training trajectory accuracy of rehabilitation robots in the presence of friction interference, a backstepping double-integral terminal sliding mode controller (BDITSMC) based on an extended state observer is proposed. An integral term was introduced into the terminal sliding surface to eliminate the arrival phase and minimize the estimation error. Meanwhile, the switching frequency in the sliding mode phase was reduced, thereby addressing the “complexity explosion” problem in inversion design. This method provides a solution to improve the trajectory tracking accuracy of upper limb rehabilitation robot controllers. In the rehabilitation trajectory tracking control, quantified by the Mean Absolute Error (MAE), the error of the backstepping integral non-singular terminal sliding mode control (BINTSMC) method is 0.0039, while the BDITSMC proposed in this paper reduces this value to 0.003, representing a 23.1% improvement.

Retraction notice: Learning model predictive controller for wheeled mobile robot with less time delay

This proposed work is presenting the approximation of higher-order (HO) underwater robotic vehicle (URV) controller with the help of multi-point matching technique by incorporating greywolf optimization algorithm (GWOA). The performance of URV system is affected by external and internal dynamics. The proper momentum of URV system is achieved by designing a controller. The URV can be effectively operated by control action of controller. The URV controller is approximated to comparatively lower-order (LO) to propose an efficient, effective and economical controller for HOURV system. The approximation is accomplished with the help of expansion parameters of HOURV controller and its desired LOURV controller. The errors between these expansion parameters of HOURV controller and its desired LOURV controller are minimized using multi-point matching. The multi-point matching is depicted in the form of objective function (OF). The constructed OF is minimized by exploiting GWOA by fulfilling the steady-state matching condition and Hurwitz stability criterion, as constraints. The effectiveness of proposed approach of multi-point matching is verified by comparing the proposed LOURV model with LOURV models obtained with the help of other approximation approaches. The applicability of proposed LOURV controller is evaluated and validated by analyzing responses and tabulated data obtained in the results. Additionally, the statistical data of performance error values (PEVs) are provided in tabulated form along with its bar plot.

It is important for legged robots to be tolerant of joint malfunctions, especially when working in extreme environments such as the surface of an extraterrestrial planet. However, it is not easy to guarantee tolerance, which requires either the robot structure and controller to be carefully designed for specific failure situations or prompt detection and diagnosis of the faults followed by quick controller reconfiguration. To solve this problem, this paper proposes a method using the automatic robot design method recently proposed by some of the authors. The core of the automatic robot design is in the method of updating the robot topology. In this paper, we combine two distinct robot topology update methods to make learning more efficient. Using a physical simulator, we simulate a situation in which several randomly selected joints malfunction. The simulation reveals that the proposed method efficiently trains a robot to walk even when some of its joints are malfunctioning without the need to switch the controller. Test results show that the obtained pair of robot structures and controllers perform well in situations not experienced during training.

To enable safe and effective human–robot collaboration (HRC) in smart manufacturing, it is critical to seamlessly integrate sensing, cognition, and prediction into the robot controller for real-time awareness, response, and communication inside a heterogeneous environment (robots, humans, and equipment). The proposed approach takes advantage of the prediction capabilities of nonlinear model predictive control (NMPC) to execute safe path planning based on feedback from a vision system. To satisfy the requirements of real-time path planning, an embedded solver based on a penalty method is applied. However, due to tight sampling times, NMPC solutions are approximate; therefore, the safety of the system cannot be guaranteed. To address this, we formulate a novel safety-critical paradigm that uses an exponential control barrier function (ECBF) as a safety filter. Several common human–robot assembly subtasks have been integrated into a real-life HRC assembly task to validate the performance of the proposed controller and to investigate whether integrating human pose prediction can help with safe and efficient collaboration. The robot uses OptiTrack cameras for perception and dynamically generates collision-free trajectories to the predicted target interactive position. Results for a number of different configurations confirm the efficiency of the proposed motion planning and execution framework, with a 23.2% reduction in execution time achieved for the HRC task compared to an implementation without human motion prediction.

This research proposes a multi-layer navigation system for indoor mobile robots when they share space with vulnerable individuals. The primary objectives are increasing or maintaining safety measures and curtailing operational costs, emphasizing reducing reliance on intricate sensor technologies and computational resources. The developed system employs a three-tiered control approach, with each layer playing a pivotal role in the navigation process. The "online" control layer integrates a human-in-the-loop strategy, where the human operator detects missing obstacles or approaching danger through a user interface and sends a trigger to the robot's controller. This trigger enables the system to estimate the coordinates of the danger and update the robot's navigation path in real time, minimizing reliance on complex sensor systems. The "semi-online" control layer generates dynamic virtual barriers to restrict the robot's navigation in specific areas during specific times. This ensures the robot avoids hazardous zones that could pose temporary risks to the human or robot. For example, areas with temporary obstructions or potential danger, such as kids' play zones or during cleaning, are temporarily restricted from the robot's path, ensuring safe navigation without relying solely on real-time sensor data. The "offline" control layer centers around the use of semantic information to control the robot's behavior according to user-defined space management and safety requirements. By leveraging Building Information Models (BIM) as digital twins, this layer combines semantic and geometric data to comprehensively understand the environment. It enables the robot to navigate according to precise user requirements, utilizing the semantic context for path planning and behavior control. This layer obviates the need for a real-time sensor mapping process, making the system more efficient and adaptable to user needs. This research represents a significant step forward in enhancing the navigational capabilities of robots within human-centric indoor environments, with a core focus on safety, adaptability, and cost-effectiveness.

ABSTRACTQuadrupedal robots are highly regarded for their superior locomotion capabilities and terrain adaptability, making them competent in a wide range of applications. For autonomous navigation, they must track upper‐level trajectories to reach designated locations with flexible obstacle avoidance. This is typically achieved by a planner, which generates a reference velocity, and a controller, which accurately tracks the velocity commands. This article proposes a learning‐based controller for quadrupedal robots that is trained in simulation and achieves accurate and robust velocity tracking in the real world. To bridge the gap betwen simulation and reality, an analytical actuator model is introduced to simulation to simulate physical actuator dynamics. We then train a control policy in simulation using Constrained Reinforcement Learning, where symmetry and smoothness constraints are incorporated into Reinforcement Learning. The symmetry constraint promotes coordinated locomotion and consistent velocity tracking performance, while the smoothness constraint reduces jerky actions and generates stable velocity performance. The proposed control policy is zero‐shot deployed on the Unitree AlienGo. Experimental results demonstrate a velocity tracking error below 0.084 m/s across the entire operational velocity range while maintaining robust locomotion on natural terrains. To further validate the controller's effectiveness, we integrate it into a pedestrian tracking framework, where it demonstrates precise trajectory following capabilities and long‐term reliability.

Robot controllers are often optimized for a single robot in a single environment. This approach proves brittle, as such a controller will often fail to produce sensible behavior for a new morphology or environment. In comparison, animal gaits are robust and versatile. By observing animals, and attempting to extract general principles of locomotion from their movement, we aim to design a single, decentralized controller applicable to diverse morphologies and environments. The controller implements the three components of (a) undulation, (b) peristalsis, and (c) leg motion, which we believe are the essential elements in most animal gaits. This work is a first step toward a general controller. Accordingly, the controller has been evaluated on a limited range of simulated centipede-like robot morphologies. The centipede is chosen as inspiration because it moves using both body contractions and legged locomotion. For a controller to work in qualitatively different settings, it must also be able to exhibit qualitatively different behaviors. We find that six different modes of locomotion emerge from our controller in response to environmental and morphological changes. We also find that different parts of the centipede model can exhibit different modes of locomotion, simultaneously, based on local morphological features. This controller can potentially aid in the design or evolution of robots, by quickly testing the potential of a morphology, or be used to get insights about underlying locomotion principles in the centipede.

Various approaches have been explored to address the path planning problem for mobile robots. However, it remains a significant challenge, particularly in environments where a multi-tasking mobile robot operates alongside stochastically moving humans. This paper focuses on path planning for a mobile robot executing multiple pickup and delivery tasks in an environment shared with humans. To plan a safe path and achieve high task success rate, a Reinforcement Learning (RL)-based double layer controller is proposed in which a double-layer learning algorithm is developed. The high-level layer integrates a Finite-State Automaton (FSA) with RL to perform global strategy learning and task-level decision-making. The low-level layer handles local path planning by incorporating a Markov Decision Process (MDP) that accounts for environmental uncertainties. We verify the proposed double layer algorithm under different configurations and evaluate its performance based on several metrics, including task success rate, reward, etc. The proposed method outperforms conventional RL in terms of reward (+63.1%) and task success rate (+113.0%). The simulation results demonstrate the effectiveness of the proposed algorithm in solving path planning problem with stochastic human uncertainties.

Human hands can precisely perform a wide range of tasks. This paper investigates key performance differences when conventional robotic hand controllers are combined with Neural Networks (NN). Tests are performed on a novel 3D printed multi-finger ambidextrous robot hand. The ambidextrous hand is actuated using pneumatic artificial muscles (PAMs) and can bend its fingers both left and right, offering full ambidextrous functionality. Force sensors are placed on the fingertips. In our control method, the grasping trajectory of each finger combines its data with that of the neighboring fingers to obtain accurate results.

Purpose The purpose of this study is to propose a trajectory generation method for 6 degrees of freedom manipulators that simultaneously optimizes all joint motions for point to point and multipoint tasks. Conventional methods treat each joint independently and synchronize to the slowest, often leading to suboptimal execution time and motion smoothness. The goal is to improve time efficiency and dynamic performance in industrial robotics by leveraging joint-level kinematic constraints. Design/methodology/approach A nonlinear optimization problem is formulated using quintic B-spline interpolation to generate smooth joint trajectories under velocity, acceleration and jerk limits. Segment durations for all joints are optimized simultaneously. The interior-point method is used to solve the problem. The approach is validated through simulation and physical execution on the Dobot Nova5 manipulator using benchmark point-to-point and multipoint motion tasks. Findings The proposed method yields trajectories up to 39% faster than conventional methods in point-to-point tasks. It produces smoother motions with lower peak and average jerk values, improving execution quality. In multipoint scenarios, the method reduces execution time by 3.26% compared to the synchronous B-spline method. Both simulation and experimental results confirm the method’s effectiveness and accuracy. Comparative analysis with recent trajectory planning methods further demonstrates the proposed method’s superior C2 continuity and smooth motion characteristics. Research limitations/implications The proposed trajectory optimization method considers only kinematic constraints, excluding dynamic factors such as torque limits, payload effects and actuator saturation. The framework assumes offline planning under fixed task conditions, limiting adaptability in dynamic environments. Obstacle avoidance is handled via predefined waypoints rather than integrated constraints. Although quintic B-splines offer a balance between smoothness and computational efficiency, alternative spline orders may improve performance in specific contexts. Future work will address dynamic modeling, real-time replanning and integration with collision-aware path planning to enhance practical applicability. Practical implications The proposed method enables smooth, time-optimal trajectory generation for robotic manipulators without synchronizing all joints to the slowest one, improving motion efficiency and productivity. Its asynchronous formulation allows better utilization of joint capabilities under kinematic constraints. By supporting both point-to-point and multipoint tasks, the approach is applicable to a wide range of industrial assembly operations. The method has been validated through hardware experiments, demonstrating consistent execution and repeatability. It can be readily integrated with existing robot controllers and extended with via point-based obstacle avoidance for constrained environments. Social implications The proposed trajectory planning method enhances the operational efficiency of robotic manipulators, contributing to increased automation in manufacturing and assembly tasks. By enabling smoother, faster and more reliable robot motion, it supports safer and more collaborative work environments, especially in human−robot shared workspaces. Improved efficiency and accuracy in industrial processes may reduce energy consumption, material waste and production costs. This advancement supports broader goals of sustainable automation and makes advanced robotics more accessible for small and medium enterprises, potentially impacting workforce roles and productivity in evolving industrial landscapes. Originality/value Unlike traditional decoupled methods, this unified framework globally optimizes all joint motions using quintic B-splines under kinematic constraints. The method is adaptable across robotic platforms and suitable for computer numerical control feedrate planning. It enables smooth and time-efficient trajectories for advanced industrial applications.

Swarm robotics holds significant potential for executing complex tasks that surpass the capabilities of individual robots. However, when robots have limited sensing abilities, their actions can be obstructed by unknown obstacles in the environment, thereby constraining this potential. To address this limitation, we propose a novel swarm robot controller that enables obstacle avoidance without sensors for detecting obstacles. Our proposed controller achieves efficient obstacle avoidance through a combination of indirect collision detection and fluidic swarm behavior using a Smoothed Particle Hydrodynamics (SPH)-based controller. The indirect collision detector utilizes only the velocity data of robots to detect collisions, thereby obtaining obstacle information without relying on specialized sensors. Furthermore, integrating collision points as virtual forces into the SPH-based controller enables effective obstacle avoidance. Our method's superiority is quantitatively validated through a comparative analysis, showcasing its significant navigation improvements under obstacle-unaware conditions. Furthermore, various investigations demonstrate the controller's utility, including pattern formation tasks and navigation tasks in dynamic obstacle environments.

This paper presents a novel, measurement-driven method for the simultaneous estimation of the User Coordinate System (UCS) and Tool Center Point (TCP) in industrial robot systems, formulated as a solution to the AX = YB calibration problem. The proposed algorithm minimizes the discrepancy between simulated and real-world robot poses by iteratively refining both coordinate transformations using external laser tracker measurements. A comparative evaluation was conducted using two ABB IRB-6640 robots and four calibration strategies, including internal three-point measurement, absolute orientation, geometrical referencing, and the proposed method. Experimental results show that the proposed approach outperforms standard industrial techniques, achieving an average translational error below 0.4 mm, compared to over 2 mm for conventional methods. The geometric method, while widely used for its simplicity, demonstrated the highest error and instability due to poor rotational alignment. The findings validate the proposed solution as an accurate, non-invasive alternative for implementing simulated programs in real production environments, particularly where direct access to robot controllers is limited.

In robotic intelligent manufacturing engineering, the machining accuracy of workpieces is directly related to the end positioning accuracy of six-degree-of-freedom serial robots, so compensating for the latter is of great significance. This article proposes a method to improve the absolute positioning accuracy of a robot by correcting the joint angles of the robot without changing the parameters of the robot controller. Firstly, establish a forward kinematics model of the robot based on the spiral theory. Then, the motion errors of each joint of the robot are measured using a laser tracker, and the RBF neural network is trained to predict the motion errors of each joint of the robot. Finally, the predicted joint motion errors are compensated for the theoretical joint angles, thereby improving the accuracy of robot end positioning. The experimental results show that the precision of the robot’s end position has been improved from 0.2456 mm to 0.0716 mm, verifying the effectiveness of this method.