This study introduces a novel and simplified method for generating the cutter flank of an internal-meshing skiving cutter for machining male screw rotors. In this approach, the cutting edge is first positioned on the conjugate surface of the rotor. The cutter flank is then generated by extruding the cutting edge along an oblique direction, which simplifies the cutter geometry and enhances the feasibility of practical tool manufacturing. Subsequently, the complete cutter structure is constructed. A mathematical model employing the proposed skiving cutter for machining a male screw rotor on a dedicated CNC machine is then presented. Simulation results obtained using VERICUT software demonstrate that the proposed cutter can successfully machine a complete rotor with high surface accuracy. Furthermore, the distribution of the working clearance angle indicates that the clearance angle remains positive along the entire cutting edge during the skiving process, thereby preventing interference between the tooth flank and the finished surface.
This contribution presents an advanced unified robot description format and its usage in the Robot Operating System (ROS). The description uses a tree structure with additional constraints included by a new “constraint”-element. The constraints are defined using two cut frames, two constraint axes and a constraint type. This enables the description of closed-loop mechanisms in arbitrary configurations. Using a new element to describe the closed loop has the advantage that the existing packages do not break when using the proposed format. These packages just ignore the closed loops and use the underlying tree structure. However, the integration of parallel robots in the ROS Framework lacks the closure of the tree structure. A Joint State Transformer package has been developed to address this issue. This package solves the constraint equations using a given set of minimal coordinates to close the open tree structure. The solver uses the Newton-Raphson algorithm to solve the constraint equations. The proposal’s results are validated using the parallel kinematics kit PARA-ENGINEER.
Efficient cleaning methods for eyeglass lenses are required for longevity of lenses and for eye hygiene. Inconsistencies in manual cleaning of eyeglass lenses using microfiber cloth has led to the development of several commercial products, including those that use ultrasonic chambers. To mimic the cleaning action of a human user, a mechanized system has been proposed in this paper, which is driven by a pair of slider-crank mechanisms powered by a DC stepper motor. It produces controlled elliptical brushing motion patterns while reducing the chances of scratches. This design also incorporates a simpler technique for pre-cleaning and aims to be environmentally friendly by reducing the need for disposable wipes. Concepts from the principles of mechanism design and kinematic evaluation were used to finalize the architecture and to obtain the parameters required for several standard shapes of the eyeglass frames. After carrying out simulations using GeoGebra, one set of parameters was used to develop a physical system to test the mechanized system.
The manual production of Banh Gio in small- and medium-scale food processing facilities is time-consuming and highly dependent on workers’ skills, leading to low productivity and inconsistent product quality. To improve production efficiency and reduce reliance on manual labor, this study proposes the design, fabrication, and experimental evaluation of an automatic Banh Gio making machine. The system integrates feeding, wrapping, and tying mechanisms into a synchronized operating cycle. Performance evaluation experiments were conducted to compare the machine with the traditional manual method. Experimental results indicate that the proposed machine achieves 5.22 times higher productivity than manual production while maintaining operational stability and product shape uniformity. This study provides a practical mechanization solution suitable for small and medium scale traditional food-processing facilities.
Rotor-bearing-coupling systems frequently suffer from complex mechanical faults, notably shaft misalignment and mass imbalance, producing strongly nonlinear vibration signatures. This paper provides a systematic overview of dynamic modeling techniques, from classical analytical theories to advanced 3D multibody and Finite Element Methods (FEM), for multi-fault diagnosis. A key novelty of this study is highlighting the critical role of high-fidelity physics-based models as robust synthetic data generators. To demonstrate this, a generalized 12-DOF 3D multi-fault model is presented, successfully capturing complex nonlinear signatures. By serving as a high-fidelity synthetic data generator, the proposed model provides rich, multi-dimensional (12-DOF) labeled datasets for extreme multi-fault scenarios, addressing the critical industrial bottleneck of data scarcity. While the current study specifically bounds its scope to the mathematical formulation and synthetic data generation phase, it establishes the essential computational groundwork for future research to train emerging hybrid AI frameworks, such as Physics-Informed Neural Networks (PINNs) and FEM-assisted Transfer Learning.
This research presents the development of a computationally efficient, C++ -based flight simulation framework designed to bridge the gap between high-fidelity professional simulators and accessible educational tools. The Cessna 172 model during cruise is described by a system of 12 nonlinear differential equations governing its flight dynamics. A comparative analysis between Euler and Fourth-Order Runge-Kutta (RK4) numerical methods ultimately proves RK4 provides superior stability and accuracy for aerospace applications. To ensure real-time performances on standard hardware, the framework implements advanced terrain generation and optimization strategies, reducing GPU load by 74
Some cases of neurosurgery require the opening of the patient’s skull through craniotomy. A section of the skull is cut off using surgical drills. Robot practice has been proposed to reduce the risk of shaking recall motion, etc. However, the use of a robotic platform imposes a difficult pre-operative procedure to program an accurate trajectory. A new method of intra-operative registration is proposed to suppress that need for pre-planning. In combination with an optronic navigation system, a trajectory can be easily defined using pinpoint designation directly on the patient skull. A specific algorithm is programed to automatically generate a trajectory that is well adapted for Remote Center of Motion (RCM) robot, which are highly recommended for this task. The algorithm is tested on an object of known geometry to evaluate its accuracy. A series of experiment are carried out on a Human skull model to verify the ability of this new method using a non-RCM robotic manipulator. Opportunely, a radar sensor is used to measure the radius variation of the skull as a preliminary specification experiment for the specification of an RCM robot for craniotomy.
Macro-mini robotic systems have recently emerged as a promising solution to combine large workspace capabilities with high dynamic performance, particularly in physical human-robot interaction (pHRI) contexts. In medical robotics, such architectures are especially relevant for ultrasound-guided percutaneous interventions, where precise and intuitive control of needle positioning is required. This paper presents a control framework for a fully actuated macro-mini robotic system dedicated to robot-assisted percutaneous procedures. The proposed system combines a KUKA iiwa collaborative robot with a lightweight planar 2-DoF mini robot and an RCM robot, enabling a clear separation between global positioning and local interactive motion. The control strategy focuses on the pre-positioning phase, allowing safe and intuitive guidance of the system close to the patient. A coupled control approach is proposed to coordinate the high-inertia macro manipulator and the low-inertia mini manipulator, ensuring smooth interaction and accurate motion execution. The effectiveness of the proposed method is validated through experiments on a real robotic platform. The results demonstrate the potential of macro-mini architectures for enhancing safety, precision, and usability in robot-assisted percutaneous interventions.
This paper presents the design of a low-cost robotic hand with an ergonomically shaped palm and opposable thumb. The hand employs a linkage-based transmission to ensure backdrivability and efficient force transfer. Kinematic models of the fingers and thumb were developed and validated through simulation. A 3D printed prototype has been built and preliminarily tested, showing accurate and repeatable movements with a broad range of motions. Results confirm the hand’s potential as an effective and affordable solution for applications requiring dexterous, reliable manipulation.
Eggplant is a widely cultivated vegetable valued for its nutritional and economic importance. However, its post-harvest processing, particularly stem removal, remains labor-intensive and highly dependent on worker skills, leading to low productivity. This study presents a prototype eggplant stem-removing machine designed to improve efficiency and reduce labor requirements. The complete processing workflow is detailed, covering stages from automatic feeding to precise positioning and stem cutting. Experimental results validate the applicability of the proposed system and show that it achieves productivity four to five times higher than conventional manual methods, with a qualified eggplant rate of approximately 95
Accurate extraction of the Time-of-Flight (TOF) is crucial for high-precision ultrasonic thickness measurement. However, the traditional fixed step-size Least Mean Square (LMS) algorithm suffers from an inherent contradiction between convergence speed and steady-state error. Furthermore, existing variable step-size algorithms rely on instantaneous errors and fixed function models. This reliance leads to poor adaptability when dealing with non-stationary echo signals. To address these issues, this paper proposes a method for estimating the TOF of ultrasonic signals based on a fuzzy variable step-size LMS algorithm. First, the proposed method abandons the conventional single-error feedback mechanism. Instead, it utilizes the local correlation coefficient error between the output and desired signals, along with its variation, as the dual-input features for the fuzzy controller. Then, a zero-order Sugeno fuzzy inference system is employed to establish the nonlinear mapping rules between the input features and the step-size factor, achieving the adaptive and dynamic adjustment of the step-size. Simulation results demonstrate that the proposed method outperforms the fixed step-size LMS, the hyperbolic tangent variable step-size LMS, and the instantaneous error-based fuzzy variable step-size LMS algorithms in overall performance. It ensures rapid convergence while significantly reducing the steady-state misalignment error, thereby providing higher measurement accuracy and better noise immunity. Ultrasonic thickness measurement experiments indicate that the proposed method yields smaller relative measurement errors for gauge blocks of various thicknesses compared to the other three LMS algorithms, with a maximum relative error of 0.710
Kresling origami structures are widely known for their complex, nonlinear dynamic behaviour, which depends on the structure’s geometric parameters. Therefore, designing Kresling origami structures with targeted dynamic responses is a significant challenge. This paper proposes a reverse design framework for Kresling structures using a neural surrogate model. We generate a training dataset from hinge-bar simulations. The natural frequencies and damping rates are extracted from free vibration data. The trained surrogate model provides rapid prediction and enables efficient geometry updates within an optimisation loop to fit specified oscillation characteristics. The proposed method significantly reduces the need for iterative nonlinear simulations and supports practical iterative design.
Due to repetitive overhead tasks and forceful exertions, the prevalence of Work-Related Musculoskeletal Disorders (WMSDs) among industrial workers is significant, with the shoulder being one of the most commonly affected regions. Wearable exoskeletons provide a promising solution to this problem. However, most exoskeletons are either purely passive or purely active, the former having limited support, while the latter being bulky and less energy efficient. This paper introduces a novel exoskeleton that integrates both active actuation and passive torque generator using a variable stiffness mechanism (VSM). The combination enhances energy efficiency and enables modulation of the impedance of the overall actuator by controlling the motor torque, thus improves exoskeleton’s long-term usability. In the paper, the design of the exoskeleton is presented. A prototype was fabricated, with control system developed. Experiments were finally conducted to demonstrate the energy efficiency of the exoskeleton.
This paper presents a trajectory generation and control strategy for a Mobile Cable-Driven Parallel Robot (Mobile CDPR), enabling it to move from an initial position to an equilateral triangular formation while the end-effector altitude is maintained. A fifth-degree polynomial trajectory is generated to coordinate the motions of the three mobile bases and the cable-winch system. The evolution of cable tensions along the trajectory is analysed through a wrench-based model to ensure that positive tensions are preserved throughout the motion. To execute the trajectory on a physical prototype, a control strategy combining feed-forward angular-velocity inputs with Proportional-Integral (PI) feedback is implemented. The PI gains are adjusted online through a gain-scheduling mechanism driven by left-right wheel error discrepancies so that traction asymmetries caused by varying cable tensions are mitigated. The approach is validated through four repeated experiments, during which tracking accuracy is preserved and tensions remain feasible across all motion segments.
Robotic-assisted gait training heavily relies on body weight support (BWS) systems to facilitate safe, early mobilization in neurorehabilitation. Despite their well-established clinical benefits, optimizing the mechanical design of these systems remains a significant challenge due to fragmented literature regarding force transparency, inertial loading, and actuation strategies. This systematic review investigates the fundamental mechanical design requirements for BWS systems, focusing on structural strength, force accuracy, mechanical transparency, and safety mechanisms. We classify current technologies into treadmill-based and overground-based architectures and compare their support mechanisms, actuation methods, and clinical utilities. Ultimately, this paper highlights critical engineering challenges, such as minimizing reflected inertia and balancing system stability with unconstrained mobility, thereby providing a comprehensive roadmap for future innovations in assistive rehabilitation robotics.
This paper presents a teleoperated manipulation framework for pedicle screw drilling in cervical spine surgery, based on a haptic master device coupled with a collaborative remote robot. The proposed approach aims to provide suitable control of the drilling motion while maintaining haptic perception of the interaction forces. First, the clinical context of cervical spine surgery is introduced, followed by a detailed description of the master–remote experimental robotic platform and its main components. The 6 Degrees of freedom (DOF) parallel haptic master device is presented, including its kinematic modeling, as well as the commercial collaborative robot used as the remote manipulator. Both systems exhibit kinematic redundancy, which is exploited within the control framework. A bilateral teleoperation control scheme is then described, detailing the control architecture, the interaction between the different subsystems, and the calibration and synchronization procedures implemented to handle the drilling task. Two experimental scenarios were defined and evaluated: a free teleoperated drilling task and an assisted teleoperated drilling task. Both strategies were tested in collaboration with medical doctors. The collected experimental data were analyzed and compared to assess drilling quality and system performance under both scenarios. Preliminary findings demonstrate the significant added value of the assisted teleoperated mode, highlighting its potential for enhanced precision and safety, and paving the way for a more comprehensive clinical validation process.
Redundant manipulators exhibit an inherently non-unique inverse kinematics (IK) relationship, where a single end-effector pose corresponds to a continuum of feasible joint configurations. This set-valued structure complicates direct pose-to-joint regression and may induce discontinuities or branch switching during trajectory execution, while conventional redundancy-resolution methods often depend on iterative, model-based computations that are costly at high control rates. This paper presents a history-conditioned neural IK formulation that exploits temporal continuity to resolve redundancy. The proposed model augments the desired six-dimensional (6D) end-effector pose with the previous joint configuration, thereby transforming the globally ambiguous inverse correspondence into a locally single-valued mapping along continuous motions. A lightweight feedforward multilayer perceptron is trained to predict the current joint state from this augmented input, enabling constant-time inference suitable for real-time deployment. The approach is validated on an 8-degree-of-freedom (DOF) redundant platform comprising a Franka Emika Panda arm mounted on a prismatic linear axis. Experiments on multi-trajectory real-robot data demonstrate accurate joint prediction and stable autoregressive rollout on held-out motions, producing smooth joint evolution and consistent task-space tracking.
This paper proposes an adaptive super-twisting sliding mode control (ASTSMC) strategy for trajectory tracking of a coaxial octorotor UAV subject to model uncertainties and external disturbances. The controller is developed based on the super-twisting algorithm to reduce chattering and is integrated with a parametric estimation mechanism for real-time disturbance compensation, enabling online estimation and rejection of lumped disturbances. Closed-loop stability and convergence of tracking errors are rigorously guaranteed through Lyapunov analysis. Simulation results under a disturbed spiral trajectory demonstrate that the proposed method shows better performance than conventional SMC and non-adaptive STSMC. Some representative results show that the average position tracking error in the x-axis is reduced by 72
Multi-modal mobile robots achieve superior maneuverability by switching among discrete locomotion or steering configurations, but existing graph-based planners frequently induce mode chattering - rapid, oscillatory switching that accelerates mechanical wear and degrades operational lifespan without meaningful improvement in path quality. This paper proposes the Graduated Dwell-Time Hysteresis (GDH) framework, which reframes dwell-time enforcement from a binary feasibility constraint into a continuous cost-regularisation mechanism compatible with optimal graph search. Formal guarantees of termination, chattering prevention, and heuristic admissibility are established in Sect. 5. Simulation on three benchmark scenarios demonstrates significant reductions in mode switches and path length versus a no-stability baseline, while a hard dwell-time constraint baseline fails due to deadlock in constrained environments.
This paper proposes a robust Semantic SLAM system designed to generate high-fidelity maps for indoor environments. Addressing the drift problem in long-term navigation, we introduce a deep learning-based loop closure module that utilizes GeM (Generalized-Mean) pooling for global place retrieval and LoFTR (Detector-Free Local Feature Matching) for precise geometric verification. This approach allows for globally consistent trajectory optimization, effectively mitigating cumulative odometry errors. Based on the optimized poses, depth measurements are fused into a TSDF volume to reconstruct a dense surface enriched with semantic information projected from 2D segmentation masks. Experimental results demonstrate that our approach effectively minimizes drift, resulting in more robust global consistency compared to standard frame-to-model tracking.
