To address the issues of insertion deflection caused by the inherent properties of standard puncture needles in clinical prostate particle implantation, the overemphasis of existing control algorithms on single puncture positioning performance at the expense of continuous puncture reliability, and the lack of precise support for soft tissue environmental stiffness, this study proposes a reliability control algorithm for puncture deflection and continuous positioning for prostate particle implantation. First, an interaction model between the puncture needle and glandular soft tissue is constructed, the deflection mechanism is analyzed, and the core role of soft tissue stiffness parameters in deflection regulation is clarified. Second, by integrating Hooke’s law with measurements from force and displacement sensors, the least squares method is adopted to achieve online estimation of human prostate soft tissue environmental stiffness, providing real-time stiffness parameters for deflection control. Finally, a closed-loop control model incorporating data acquisition, deviation identification and instruction feedback is constructed, the needle posture is adjusted via spatial transformation of the puncture robotic arm, a particle filter algorithm is adopted to evaluate positioning reliability in real time, and the spatial adjustment of the needle tip is derived to ensure continuous puncture accuracy. Comparative experiments on prostate bionic soft tissues demonstrate that the proposed online stiffness estimation method takes approximately 3 s and can rapidly and synchronously track dynamic stiffness variations during puncture, and that the deflection and continuous puncture control method is effective and reliable, achieving a target positioning overlap rate of 95.6%. This algorithm ensures the therapeutic efficacy of prostate particle implantation and provides a phased theoretical reference for research on precision control systems for puncture robots.
Accurate localization of unmanned aerial vehicles (UAVs) is essential for applications such as structural health monitoring, especially in environments where Global Positioning System (GPS) signals are denied or unreliable, like indoor spaces, tunnels, urban canyons, or areas beneath large structures. To address this challenge, we propose Cross-Fusion, a novel method for real-time UAV localization that integrates data from a 3D Light Detection and Ranging (LiDAR) and a monocular camera. A key contribution is its cross-session fusion strategy, which integrates visual and geometric information collected from multiple agents during routine baseline surveys to improve localization consistency and map completeness. The system employs LiDAR-based odometry for motion tracking and image-based feature matching via a single red-green-blue (RGB) camera to correct drift and improve accuracy. Unlike visual-inertial systems, Cross-Fusion maintains a simple sensor setup and avoids the complexity of stereo or global shutter configurations. Experimental results demonstrate that Cross-Fusion achieves localization accuracy comparable to GPS-based methods and performs reliably in challenging feature-sparse environments.
Water sample and monitoring is essential for protecting aquatic ecosystems and detecting environmental pollution. This paper presents the design and experimental validation of a bio-inspired miniature submarine for low-cost water sample and monitoring. Inspired by the jet propulsion mechanism of squids, the proposed system employs pump-driven water jets for propulsion and steering, combined with a pump-based buoyancy control mechanism that enables both depth regulation and water sampling. The vehicle integrates low-cost, commercially available components including an ESP32 microcontroller, IMU, pressure sensor, GPS receiver, and LoRa communication module. The complete system can be constructed at a hardware cost of approximately $122.5, making it suitable for educational and environmental monitoring applications. Experimental validation was conducted through pool tests and field trials in a lake. During a 360 degrees rotation test, roll and pitch deviations remained within +/- 2 degrees and +/- 1.5 degrees, respectively, demonstrating stable attitude control. Steering experiments showed a heading step response with approximately 2 s rise time and 5 s settling time. Depth control experiments achieved a target depth of 2.5 m with steady-state error within +/- 0.1 m. Field experiments further demonstrated reliable navigation and successful water sampling operations. The results confirm that the proposed platform provides a compact, stable, and cost-effective solution for small-scale environmental monitoring in shallow-water environments.
Achieving precise robotic manipulation is difficult when objects are unknown or not pretrained, and the environment is unpredictable. Traditional cameras often struggle with depth perception and lag when a robot gets close to its target. To address these limitations, this study explores the design and implementation of an image-based visual servoing (IBVS) system for a 5-degree-of-freedom (DOF) robotic arm mounted with an eye-in-hand camera and sonar sensor. The proposed system enables real-time object tracking and autonomous grasp using fusion of visual and depth feedback in the Robot Operating System (ROS) 2 and MoveIt Servo frameworks. A Channel and Spatial Reliability Tracker (CSRT)-based tracking algorithm is applied for precise object localization and consistent performance after a comparative analysis with some other tracking algorithms for object dynamic movement up to 30 rpm. The depth estimation using the sonar sensor provides an average error of 1.2 cm in the 5-30 cm working range to provide accurate reconstruction of the object's 3D location. The MoveIt Servo controller generates joint-space trajectories with smooth velocity changes. Experimental trials with 40 grasping tasks had a total success rate of 80% in manipulation, with consistent performance for objects of different sizes and shapes. Results above confirm the correctness of the presented method in providing adaptive robot control and reducing human operators' manual operation effort. This hybrid feedback approach provides a reliable foundation for real-time visual servoing, overcoming the latency and precision constraints of traditional vision-only methods.
The bogie of freight railcars is a critical component affecting train safety and maintenance automation. Traditional pose localization and bolster tilt detection methods rely heavily on manual measurements and laser ranging, which suffer from high environmental sensitivity, limited accuracy, and poor real-time performance. To address these limitations, this study proposes a visual pose detection system for bogie bolster tilt based on deep information fusion. The system includes hardware design using depth cameras, spatial alignment of RGB and depth data, and temporal filtering for depth smoothing. An adaptive feature selection module and coordinate attention (CA) mechanism are incorporated into the detection network to enhance feature extraction. Experimental results show that the system achieves pose detection errors within +/- 4 mm for depth, +/- 3 mm for positional offset, and +/- 0.5 degrees for rotation. Further improvements using adaptive filtering, optimized depth map matching, and compensation function refinement reduce depth and offset errors to within +/- 2 mm, while maintaining rotation error within +/- 0.5 degrees. The results validate the proposed method's effectiveness and its potential for intelligent bogie inspection applications.
This study addresses the challenge of path planning in mobile robots, that requires efficient navigation in complex environments. Traditional approaches often struggle to meet the increasing demands of modern multi-robot systems operating in dynamic environments. To address these limitations, this study proposes an improved path planning technique by combining the probabilistic roadmap (PRM) with the genetic algorithm (GA), forming a hybrid PRM-GA approach designed to optimize the routes of mobile robots. Experiments were carried out for scenarios involving 2, 3, and 9 robots to analyze the performance of the proposed method under increasing complexity. The proposed PRM-GA method was compared with widely used path planning algorithms including A & lowast;, Rapidly exploring random tree (RRT), and conventional PRM. Performance of each method was evaluated focusing on path efficiency and energy consumption. The enhanced fitness function within the GA evaluates robot paths based not only on distance but also on smoothness and turn count, promoting routes with fewer directional changes. The proposed PRM-GA method reduces robot energy consumption while improving navigation efficiency. Experimental results demonstrate that the PRM-GA hybrid method outperforms A & lowast;, RRT, and PRM by encouraging smoother paths with fewer turns, thereby enhancing the operational efficiency of multi-robot systems. The effectiveness of the proposed approach highlights its potential for practical applications in sectors where efficient mobile robot navigation is essential.
The primary objective of a soft robotic finger is to reproduce the functional versatility of the human hand by incorporating intrinsic flexibility. By mimicking the human finger’s morphology, structural organization, and kinematic behavior, the system enables anthropomorphic motion that provides enhanced adaptability and dexterity. This study introduces the design and evaluation of an underactuated, anthropomorphic finger prototype actuated by a pneumatic muscle. The natural compliance of the actuator allows the finger to adapt its shape to the contours of the grasped object. As a result, the proposed design supports both biomimetic motion in free space and adaptive grasping actions during interaction with a wide range of objects.
BackgroundEffective preparation is essential before deploying robot-assisted surgery (RAS) in the operating room (OR).MethodsIn this study, we present the Robossis surgical simulator (RSS), a haptic, fluoroscopy-guided simulator designed for femur-fracture surgery. Twenty medical students were randomly assigned to either the virtual reality (VR) head-mounted display (HMD) or non-VR (i.e., on-screen) groups.ResultsBoth groups showed significant pre-to-post gains across metrics (p < 0.05), with shorter times (on-screen approximate to -39%, HMD-VR approximate to -40%), improved accuracy (on-screen approximate to -77%; HMD-VR approximate to -52%), reduced path length (on-screen approximate to -44%; HMD-VR approximate to -46%), and fewer collisions (on-screen approximate to -55%; HMD-VR approximate to -63%). Workload decreased for most domains in both groups; physical demand did not change significantly in the on-screen group (p = 0.1821). In the between-group post-training, the only difference was in translational accuracy (p = 0.041) for the on-screen group.ConclusionThese findings highlight the learning curve associated with VR training for RAS.
Wheel-legged robots offer high mobility on flat terrain and adaptability in complex environments, yet achieving stable motion under uncertain ground conditions and during height variations poses significant control challenges. This study presents a control framework for a bipedal wheel-legged robot (BWLR) to enable stable locomotion and posture transitions on uneven and inclined terrains. An adaptive Fuzzy-linear quadratic regulator (Fuzzy-LQR) controller is presented, where a fuzzy supervisory layer dynamically tunes the LQR gains and estimates the center of mass (CoM) using intermediate postures. Additionally, a proportional-derivative (PD)-based hip stabilization controller is integrated to regulate the roll angle via coordinated hip joint actuation, enhancing balance on uneven surfaces. The framework is implemented in a Robot Operating System (ROS)2-based architecture with torque-level joint control. Its effectiveness is validated through high-fidelity simulations in ROS2-Gazebo and real-world experiments on a physical BWLR prototype, including detailed aspects of multiplugin integration and torque-feedback hip control. Results demonstrate robust height adaptation during locomotion and stable balance on inclined terrain, highlighting the adaptability of the combined Fuzzy-LQR and PD approach for wheel-legged robotic systems.
The integration of multimodal sensor data is critical for developing intelligent systems capable of extracting complex, insightful information. We introduce a robust platform for high-fidelity and real-time hand posture replication, as well as object classification, centered on a novel tactile-sensing glove. The glove design incorporates 10 flex sensors spanning the PIP and MCP joints, complemented by two MPU6050 IMUs strategically mounted on the thumb’s TM and the hand’s dorsal CMC bones for comprehensive motion capture. We leverage this rich sensory data to drive a digital twin—constructed in Blender and visualized in the Unity3D engine—achieving precise real-time hand posture reproduction, necessary for visual feedback and sensor anomaly detection during sensor calibration and data collection by the operator. Furthermore, we evaluate the glove’s classification capability on a custom dataset of 15 objects. Through a grid search optimization, we trained One-dimensional Convolutional Neural Network (CNN-1D), Long-Short Term Memory (LSTM), and Temporal Convolutional Network (TCN) architectures. The models achieved classification testing accuracies confidence intervals of [95.40%, 96.50%], [92.55%, 93.84%], and [93.47%, 94.76%], respectively, validating the high-performance and utility of our multimodal sensing approach.
Trajectory planning and control of bionic dual-arm underwater robotic manipulators (BDA-URMs) are highly challenging due to the complex interactions between the manipulator's base, its dual arms, and the nonlinear underwater dynamics. This study proposes a novel hybrid controller to enhance trajectory tracking and control of a BDA-URM. The proposed hybrid control strategies integrate a proportional-integral-derivative (PID) controller and overwhelming control, which are tuned using various methods, including Ziegler-Nichols (Z-N), genetic algorithm (GA), ant colony optimization (ACO), and particle swarm optimization (PSO) for generating precise motions. A comprehensive Simulink model is developed, incorporating the bionic dual-arm, buoyancy, and hydrodynamic forces to analyze system dynamics. The PSO-based tuning of the hybrid controller demonstrated superior trajectory tracking performance compared to other methods. The controller achieved significant reductions in the integral of time multiplied by absolute error (ITAE) for different joints: link1-joint1 (Arm1), link1-joint2 (Arm2), link2-joint3 (Arm1), and link2-joint4 (Arm2), with improvements of 92.44%, 90.6%, 87.46%, and 89.21%, respectively. The proposed hybrid control method achieved the required stability with an exceptionally fast response time of 0.0089 s, making it highly effective for trajectory control in underwater manipulators.
BackgroundThe aim of this study lies in the development of a hybrid Cartesian robot for knee joint surgeries, to eliminate the shortcomings of existing surgical robots.MethodsTo develop a hybrid Cartesian robot, we used a combination of a Cartesian manipulator and a serial manipulator, both with three degrees of freedom (DOF).ResultsA 3D model of a hybrid Cartesian robot for knee joint surgery is developed, and a prototype is designed in a simplified form based on this model. The experimental tests of the 6DOF prototype of the hybrid Cartesian robot used for knee joint surgery confirmed its good performance when cutting hard cow bone.ConclusionTo eliminate the disadvantages in surgical robots used for knee joint operations, a hybrid Cartesian robot has been developed. The hybrid Cartesian robot has the strengths of parallel and serial robots, potentially eliminating the main disadvantage of their designs.
Aiming at the requirement that the brush head must maintain a specific distance from the curved surface during the cleaning operation of complex surfaces such as vases and cultural relics, this paper proposes a fixed-distance control method based on distance feedback from three non-collinear points at the end effector. The method collects distance information through three laser ranging sensors arranged in a triangular distribution, calculates the position deviation by averaging the three-point distances and dynamically adjusts the position of the end effector to maintain the preset working distance. To improve the ranging stability and accuracy of the sensors, moving average, weighted moving average, median filtering, Kalman filtering, and a combined filtering algorithm of moving average + weighted moving average are introduced. Comparative experiments verify that the Kalman filtering algorithm under a sampling period of 180 ms is optimal, reducing the root mean square error (RMSE) of distance measurement to 0.11 mm. To verify the effectiveness of the method, a compact cleaning robotic arm is designed, which realizes the combined motion of low-precision large-angle movement of the arm and high-precision small displacement of the end effector through a single-chip microcomputer and a bus servo control module. Experimental results show that the coordinate deviations of the system are controlled within 1.5 and 2.02 mm in fixed-distance vertical and horizontal motion experiments, respectively; in the irregular curved surface experiment, the maximum deviation is 1.35 mm and the RMSE is 0.61 mm, meeting the precision requirements for complex surface cleaning.
Tiny flying insects rely heavily on optical flow for landings and navigation. By maintaining a constant optical flow divergence, they can approach targets or obstacles with an exponential decay of both relative distance and velocity. Previous studies have shown that Micro Air Vehicles (MAVs) leverage this control strategy for efficient landings and safe obstacle approaches. However, a key question remains in both biological systems and MAVs: when is the moment to extend an insect’s legs or turn off the motor for landing, or to stop in front of obstacles? To address this, we propose a method that utilizes visual appearance cues to make these decisions. Our approach extracts visual features using texton distribution and employs the chi-square test with data-driven adaptive thresholds to detect when an MAVs should halt and what visual appearance characterizes an obstacle. Several flight tests with varying flow divergence setpoints validate that this method ensures smooth obstacle approaches while effectively determining a safe stopping point. Additionally, the results show that MAVs can segment obstacles from a distance using the detection results. A potential application of this approach is in swarm navigation, where sharing minimal, onboard-processed visual cues allows MAVs to efficiently predict obstacle size in advance, which facilitates coordinated path planning and collision avoidance. By leveraging cooperative perception strategies, this could significantly enhance the autonomy and efficiency of MAVs swarms in complex environments.
With the increasing global incidence of strokes, the demand for the development of rehabilitation robots for poststroke patients is very high. However, the development of rehabilitation robots for poststroke patients is a complex issue. To effectively engage in research in this field, one needs to be equipped with a vast amount of knowledge related to biomechanics, medical treatment, robotic systems, and control systems. For new developers entering this field, it will be a significant challenge, requiring an understanding of various interdisciplinary areas. Therefore, the aim of this research is to systematically evaluate the interdisciplinary knowledge areas used in the development of lower limb rehabilitation robots for poststroke patients. This study will assess several fundamental areas of knowledge, including human biomechanics, anthropometry, stroke pathology, treatment methods, lower limb rehabilitation robotic systems, safety issues, and clinical trials for the system. Additionally, this research will discuss some potential development directions in this field. Through this study, essential knowledge content necessary for the development of lower limb rehabilitation robots will be provided.
Fiber-reinforced elastomeric enclosures are widely used due to their lightweight, low-cost, and compliant structure. However, these actuators produce only a single motion when pressurized, based on their fixed, embedded fibers. In contrast, active fiber-reinforced elastomeric enclosures (AFREEs) use active fiber-constraint layers to create different motions, enabling versatility. In this paper, we introduced and verified a static fiber model to optimize AFREEs. Next, we created AFREEs that use both embedded and active constraints and measured their twist. We implemented two machine learning (ML) models, a support vector machine (SVM) and a fully connected neural network (FCNN), to create a forward and inverse static response model. The static fiber model yielded an average error of 2.5%. Our AFREE twisted up to 110 degrees. There were strong fits for the forward static response model (SVM: R2 = 0.9917; FCNN: R2 = 0.9901) and moderate fits for the inverse static response model (SVM: R2 = 0.6985; FCNN: R2 = 0.7656). Our results show the strength of data-driven AFREE modeling, while laying the groundwork for customizable AFREE units.
This article proposes a novel four-wheeled mobile robot with footprint reconfiguration capability. A variable footprint mobile robot is a type of robot that can reconfigure (reduce or expand) its footprint according to the dimensional constraints of the environment. In the proposed mobile robot, footprint reconfiguration is made possible through a novel transformable chassis based on a modified version of Hart’s A-frame and a slider-crank type actuation mechanism. Theoretically, the proposed transformable chassis can grant footprint reductions up to 51%. The mobile robot comprises four main modules: transformable chassis, actuation mechanism, suspension system to reduce ground-induced vibrations, and wheel drive system to facilitate locomotion. This modular design promotes customizability, enabling creation of different embodiments. The developed prototype of the mobile robot demonstrated a maximum footprint reduction of 40.7%. During the implementation, it demonstrated semi-autonomous navigation using a combination of a simultaneous localization and mapping (SLAM) algorithm, a dynamic window approach (DWA) local planner, and a Dijkstra’s algorithm-based global planner, navigating through an arena while utilizing its footprint variation capability to travel through narrow regions.
Soft robotic grippers show exceptional promise for handling delicate, irregular objects, but face a tradeoff between high-precision manipulation and heavy-payload capability. As per the authors' knowledge, there is a lack of existing systems that offer hardware-level modularity for on-demand reconfiguration. To overcome the existing gap, the article presents a modular and scalable pneumatic soft gripper with quick-swappable silicone fingers. The proposed methodology combines (1) magnetically coupled modular fingers (swapped in <5 s), (2) parametrically scaled PneuNet actuators (S -> L sizes), and (3) self-sealing pneumatic interfaces. The experiments show linear payload scaling from 0.32 kg (60 mm fingers) to 4.23 kg (150 mm fingers) (R-2 = 0.98), 96.4% grasp success across fragile, irregular, and heavy objects, and 100% swap reliability over 200 cycles. This work enables a single gripper to adapt from micro-precision to heavy-payload tasks, advancing soft robotics toward cross-domain automation.
With the advancement of automation technology, flexible arm robots are increasingly widely used in industrial, medical, and other fields, especially in complex and uncertain environments, showing great potential. However, flexible arm robots face many challenges in path planning (PP) and motion control, such as dynamic deformation and nonlinear characteristics affecting motion accuracy. The existing PP algorithms and control methods have shortcomings in terms of dynamic obstacle adaptability and energy consumption. Therefore, a flexible arm robot tracking algorithm combining improved ant colony algorithm and dynamic window approach (DWA) is proposed, which improves trajectory tracking performance through a dual closed-loop control (CLC) strategy and an extended state observer (ESO). The experimental results showed that in the PP task, the success rate of the raised tracking algorithm in dynamic obstacle environments decreased from 98.66% to 73.69%, which was higher than the comparison algorithms. In the trajectory tracking task, the proposed control strategy reduced the average error on the X-axis and Y-axis to 1.34 and 1.02 cm, respectively, which was much better than traditional control strategies. In addition, the introduction of an ESO improved the compensation capability for external disturbances, reducing trajectory deviation by over 67%. The research results demonstrate the effectiveness and superiority of the raised tracking algorithm in dynamic and complex environments, providing a new solution for PP and control of flexible arm robots, which has important theoretical value and practical significance.
With growing demands for machining large-scale complex components in aerospace and marine engineering, conventional processing methods face bottlenecks, including insufficient rigidity, error accumulation, and limited flexibility. Parallel mechanisms (PMs), with their high stiffness and dynamic pose reconfiguration capability, provide critical kinematic platforms for in situ machining. However, their performance optimization has long been constrained by multiobjective coupling and significant dimensional heterogeneity. This study proposes a collaborative optimization strategy integrating radial basis function neural network (RBFNN) and multiobjective particle swarm optimization (MOPSO), applied to a novel decoupled dual-platform PM. Based on finite instantaneous screw (FIS) theory, we establish the topological kinematic model and performance quantification index system for the decoupled dual-platform mechanism. Optimal Latin hypercube sampling (Opt LHS), Z-score normalization, and RBFNN are employed to eliminate dimensional heterogeneity among multiple objectives, constructing a high-precision surrogate model that replaces traditional analytical optimization processes. The minimum distance method is adopted to select Pareto-optimal solutions, achieving multiobjective collaborative optimization in a five-dimensional design space encompassing full-orientation workspace and mechanical transmission efficiency. Experimental results demonstrate that the proposed method achieves: 6.64% expansion in full-orientation workspace, 6.82% improvement in dexterity, 14.10% enhancement in antideformation capacity, 4.49% increase in global minimum load-bearing capability, and 4.90% improvement in global mechanical transmission performance. This work presents a theoretically innovative and engineering-feasible solution for topological optimization of high-precision machining equipment.
