Engineering

Abstract: This study proposes a geometric procedure for robust controller tuning under parametric uncertainty, based on root-contour analysis of the closed-loop control system. For a fixed candidate controller tuning, the set of possible pole locations induced by the admissible variations of the control plant parameters is constructed. Robust admissibility is formulated as a geometric set-inclusion problem, requiring this set to remain inside a prescribed dynamic performance region in the complex s-plane. A distinction is introduced between nominal admissibility, robust stability, and robust admissibility, showing that stability over the entire uncertainty set is not sufficient to guarantee the desired dynamic performance. To quantify the root contours, several indices are defined, including the dispersion along the real and imaginary axes, the maximum pole displacement with respect to the nominal pole locations, and the geometric margin to the boundary of the performance region. The procedure is applied to the selection and verification of PI controller tunings for an uncertain single-input single-output (SISO) control system and is further validated through examples with different structures of parametric uncertainty, including a system with a single uncertain parameter and a PID-controlled system with several uncertain control plant parameters. The results show that root-contour analysis can distinguish tunings that are only robustly stable from tunings that preserve the prescribed dynamic performance over the entire uncertainty set. Thus, the method can be used as a practical tool for the diagnosis, comparison, and selection of controller tunings under parametric uncertainty.

Abstract: Rising complexity in cyber-physical systems development exposes challenges in the consistent and reusable specification of graphical domain-specific languages (DSLs). Despite the benefits of model-based systems engineering (MBSE), the absence of a standardized, lifecycle-wide specification process results in semantic inconsistencies, tool dependence, and limited interoperability. While our previous work has addressed individual stages of DSL definition, a comprehensive, standards-based process integrating these stages remains missing. Building on these foundations, this paper introduces a unified language specification process for graphical DSLs grounded in established standards---the Meta-Object Facility (MOF), Unified Modeling Language (UML), Web Ontology Language (OWL), and Resource Description Framework (RDF). The process integrates three core artifacts: a tool-independent ontology capturing domain semantics, a MOF-conformant metamodel unifying abstract syntax, semantics, and concrete syntax, and a UML-profile-based implementation. To support and exemplify this process, a prototypical toolchain is introduced that enables automated transformations between these artifacts, thereby facilitating the consistent propagation of semantics from ontology to implementation. The applicability of the proposed process is demonstrated through both a top-down automotive case and a bottom-up cybersecurity DSL, illustrating its cross-domain generalizability. By explicitly structuring and connecting ontology, metamodel, and implementation, this work contributes a semantically consistent, machine-interpretable, and tool-independent specification process for graphical DSLs in MBSE.

Abstract: This paper addresses the robust trajectory tracking problem of an Unmanned Aerial Vehicle (UAV) equipped with a 2-DOF manipulator, designed for fast aerial manipulation of varying payloads. To overcome the high computational cost and adaptability limitations of traditional model-based controllers, this work introduces a novel hybrid gain-scheduling framework that shifts the computational complexity to the pre-flight phase. The approach utilizes an approximate inverse dynamics linearization, based on fixed nominal models, which transforms the complex nonlinear system into a simple linear plant with bounded, structured uncertainties. The entire configuration space, including manipulator states and a range of payload properties, is partitioned into dynamically similar regions using K-Means clustering. For each local region, a dedicated robust PD controller is designed using a multi-objective Genetic Algorithm (GA). This framework also successfully implements a gain interpolation technique to mitigate the potential for abrupt control actions. Simulation results validate the controller’s ability to maintain high-precision tracking during fast maneuvers and payload switching, confirming the robustness and adaptability of the offline-tuned design.

Abstract: During the measurement process, load cells are susceptible to temperature variations, which can significantly degrade measurement accuracy. To address this problem, this paper presents a temperature compensation method based on an improved neural network. First, the mechanism of sensor temperature drift is analyzed from a thermodynamic perspective. Subsequently, an Improved Honey Badger Algorithm (IHBA) is developed to optimize the initial weights and biases of a Back-Propagation (BP) neural network, aiming to enhance global search capability and convergence stability. To validate the proposed method, a dedicated calibration experimental system was constructed, and temperature-dependent output data were collected over a range of 0 °C to 60 °C. Comparative experiments with conventional methods, including IMA-BP, PSO-BP, standard BP, and polynomial fitting, were conducted. In addition, an ablation study was performed to verify the effectiveness of the proposed improvements. The results demonstrate that the IHBA-BP model achieves superior compensation performance. The temperature drift coefficient and sensitivity temperature coefficient are reduced by 86.6% and 95.86%, respectively. The proposed method shows strong potential for improving measurement accuracy of load cells under varying temperature conditions and provides a practical solution for industrial sensor calibration applications.

Abstract: This work provides an experimental comparison between classical PID, analytically compensated PID, and fuzzy control applied to the speed control of a rover actuator based on a permanent magnet DC motor. Unlike most studies, which focus on classical metrics such as transient response and steady-state error, this work incorporates kinematic indicators such as acceleration and jerk to characterize the dynamic effort applied to the actuator. The results indicate that the fuzzy controller achieves the fastest transient response and the best disturbance rejection, although at the cost of an IAJ 2.378 times higher than that of the classical PID and a peak jerk 79.36% higher under nominal conditions. The classical PID exhibits the smoothest kinematic profile under nominal operation, but under disturbances it generates jerk peaks 2.39 times higher than the fuzzy controller and an IAJ 1.67 times higher than the compensated PID, evidencing its inadequacy under variable loads. The compensated PID achieves the lowest cumulative IAJ under disturbance, outperforming the fuzzy controller by 6.7%, and provides the best overall balance between response speed, disturbance rejection, and cumulative mechanical wear.

Abstract: Emerging technologies and cyber-physical systems have led to the development of complex mathematical models described by differential equations with multiple fractional orders. In this regard, this paper investigates the stability of control systems for this class of models, defined by state equations with multiple fractional orders ranging between 0 and 1. Matrix criteria and comparison principle for linear and nonlinear autonomous systems of different fractional orders are developed based on generalized Lyapunov functions for differential equations with multi-order fractional exponents. The results are extended to non-autonomous linear or with nonlinear components systems of different fractional orders. The application of the Yakubovich-Kalman-Popov lemma, adapted for this class of systems, allows us to obtain new stability criteria presented as frequency criteria and represented graphically by familiar frequency plots similar those of the Nyquist or Popov type. Numerical applications illustrate these results such as models of complex human-machine systems described by state equations of multivariable fractional orders. An analysis of the advantages of the proposed methods compared to procedures and techniques used in other papers regarding the study of multi-order fractional exponent systems is presented. It is demonstrated that the proposed methods minimize the computational effort required for stability criteria.

Abstract: An approach for the control of linear control systems with a single time-delay is proposed. The main contribution is the inclusion of a symmetric-injection virtual reference trajectory into the controller to render stability robustness of single-delay linear control systems. The dynamics of the virtual trajectory is included into the closed-loop dynamics allowing theoretical computation of the critical time-delay before losing stability. Moreover, an energy-based symmetry interpretation of the proposed approach is drawn. Numerical simulations considering stable and unstable linear systems are shown, and experiments to control a DC-motor with time-delay measurements validate our proposal.

Abstract: Drill bits can be one of the toughest components to maintain when working with CNC systems because of their unique geometries and slow wear of the tools themselves. When measuring wear on drill bits, it’s important to consider the impact tool wear can have on the drill's accuracy, the smoothness of the surfaces created, and the overall efficiency of the machining process. The wear of drill bits is a common occurrence and a normal part of the machining process. This paper seeks to address these challenges by implementing a classification framework for tool wear in CNC drill bits that utilises the Synchrosqueezed Wavelet Transform (SSWT) and the Vision Transformer (ViT). During controlled drilling experiments, Acoustic Emission (AE) signals were captured for each of the following tool conditions: Healthy Tool (HT), Low Wear (LW), Medium Wear (MW), and Severe Wear (SW). In this study, the wear of drill bits was measured and created artificially, with Electrochemical Machining (ECM) for drill bits of sizes 3.0 mm, 3.2 mm, 3.4 mm, 3.6 mm, and 3.8 mm. A system by National Instruments (NI) was used for data acquisition, and LabVIEW was used to acquire a set of data with high resolution and time-frequency representation developed with the SSWT method, which is designed for drill bit wear measurement. These features were captured in the SSWT time-frequency maps, which were used as input to a Vision Transformer that enables efficient capture of global relationships in the time–frequency domain. Unlike traditional convolution-based methods, the proposed transformer-based framework allows for automated multi-domain fusion and feature learning. During experiments with 10-fold cross-validation, the proposed SSWT-ViT framework demonstrated reliable generalisation, strong robustness, and high classification accuracy across varying wear states. Thus, the proposed method is appropriate for intelligent real-time monitoring of CNC drill bit conditions in an industrial setting.

Abstract: The paper proposes an implementation of Kolmogorov-Arnold networks (KANs) for the purpose of dynamic proportional-integral-derivative (PID) control tuning in first- and second-order linear systems under noisy and time-varying reference conditions. Toy da-tasets, based on instantaneous system error, output and reference trajectory, are used for training the networks and comparing KANs results over a performance of: i) a PID with fixed coefficients, taken from MATLAB’s Simulink PID Autotune; ii) an MLP based neural network (NN), trained on the same datasets; iii) a traditional adaptive PID scheme with gain scheduling; iv) an LMS-based online tuning approach. Results show that KANs out-perform MLPs and LMS even with less optimized datasets under noisy and quick-changing conditions and perform on par with methods, such as gain scheduling, while allowing for more flexibility and easier setup.

Abstract: This paper presents a Digital Twin (DT)-based framework for the control, monitoring, and intelligent optimization of an Assembly/Disassembly/Repair Mechatronic Production Line (A/D/R MPL), developed as a laboratory platform aligned with Industry/Education 4.0/5.0 paradigms. The A/D/R MPL is assisted by two complementary cyber–physical robotic systems: an Assembly/Disassembly/Replacement Cyber–Physical Robotic System (A/D/R CPRS), and a Mobile Cyber–Physical Robotic System (MCPRS), enabling both fixed and mobile intelligent operations. The CPRS is equipped with an industrial robotic manipulator (IRM) responsible for A/D/R tasks, while the A/D Mechatronic Line (A/D ML) consists of seven interconnected workstations (WS1–WS7) dedicated to storage, transport, quality control, and final product handling. MCPRS includes a wheeled mobile robot (WMR), carrying a robotic manipulator (RM) and Mobile Visual Servoing System (MVSS). Each workstation is connected to a local slave programmable logic controller (PLC), which communicates via Profibus with a master PLC located at the CPRS level. Additional communication infrastructures include LAN Profinet and LAN Ethernet for local integration, and WAN Ethernet connectivity enabled through OPC-UA, ensuring interoperability, scalability, and remote accessibility. Virtual environment supports task planning through Augmented Reality (AR) and real-time monitoring through Virtual Reality (VR). The system behavior is modelled with synchronized hybrid Petri nets (SHPNs) which describe the discrete and hybrid dynamics of A/D/R processes. Artificial Intelligence (AI) techniques are integrated into the DT framework for optimal task scheduling and adaptive decision-making. As a laboratory-scale implementation, the proposed system provides a comprehensive platform for experimentation, validation, and education. It supports Education 4.0/5.0 objectives by facilitating hands-on learning, human–machine interaction, and the integration of emerging technologies such as AI, digital twins, AR/VR, and cyber–physical systems. At the same time, it embodies Industry 4.0/5.0 principles, including interoperability, decentralization, sustainability, robustness, and human-centric design.

Abstract: Leveraging its exceptional ultra-low altitude flight capability and high economic effi-ciency, the unnamed Wing-in-Ground (WIG) craft offers unique advantages in mari-time missions such as island patrol and rapid replenishment. However, the path plan-ning for unnamed WIG crafts faces the dual challenge of precise obstacle avoidance and ultra-low altitude maintenance, due to the obstacle distribution in island regions and the altitude window constraints inherent to ground effect flight. To address this, this paper proposes a path planning method based on an improved hybrid Sparrow Search Algorithm and Grey Wolf Optimizer. This method integrates the swarm intel-ligence of the Sparrow Search Algorithm, employs a self-destruction mechanism to es-cape local optima. Furthermore, it combines the hierarchical guidance of the Grey Wolf Optimizer to enhance convergence accuracy. The algorithm integrates ground-effect maintenance constraints and a reef threat model, and smooths the final path using cubic B-spline curves. Simulation results demonstrate that the proposed algorithm outperforms the standard Sparrow Search Algorithm, Grey Wolf Optimizer, and Particle Swarm Optimization in terms of convergence speed, optimization accu-racy, and obstacle avoidance success rate. It is capable of generating a feasible, safe, and smooth path, thereby supporting the autonomous navigation of unmanned WIG crafts in island reef waters.

Abstract: As the era of Industry 4.0 (i.e., the fourth-generation industrial revolution) develops, machine tools in particular are becoming interconnected, forming a collaborative community in smart factories. “Smart manufacturing” is becoming the norm, in a world where intelligent machines, systems, and networks are able to exchange information between them and respond independently, with autonomy to information, with the goal to manage industrial production processes. An important challenge is the transformation of traditional machines into intelligent machines, respectively intelligent or smart spindle. The purpose of this paper is to analyze the spindle using the intelligent models in in-situ conditions. The historical evolution, recent challenges and future trends of machine tool spindles were analyzed, noting that further development would be necessary to enable sensor/AI module integration to make the spindle unit an inherent quality assurance system. This study proposes a deep learning-based approach to spindle health monitoring based on multi sensor vibration signal analysis. The two proposed AI methods is based on the analysis of acceleration and synchronous envelope vibrations by demodulating the signal based on the Hilbert transform to identify critical bearing defects and specific defects at high frequencies.

Abstract: The control of highly nonlinear, open-loop unstable dynamics is a prevalent engineering challenge, often benchmarked through Magnetic Levitation (Maglev) systems. While continuous-time adaptive neural networks are commonly used to reject disturbances, their direct digital implementation often induces closed-loop instability due to unaccounted sampling effects. To address this, this paper proposes a Discrete-Time Fourier Series Neural Network (FSNN) control architecture for nonlinear single-input single-output (SISO) systems that can be transformed into the Brunovsky canonical form. The parameter adaptation laws are synthesized strictly in the discrete-time domain using Lyapunov stability theory. This approach yields an explicit upper bound for the digital sampling period, ensuring a proper implementation. Furthermore, it guarantees the Uniform Ultimate Boundedness (UUB) of the tracking error in the presence of bounded unmodeled dynamics and periodic disturbances. Numerical simulations of Maglev dynamics validate the theoretical bounds, demonstrating that the FSNN controller achieves rapid learning and generates a smooth control effort, offering a robust and practical framework for digital control.

Abstract: This study presents a hybrid controller that integrates fuzzy logic control and the incremental conductance method. This controller optimizes maximum power point tracking in a 330 W photovoltaic system by designing a step-down DC-DC converter. The study evaluates the impact of the number of membership functions (three, five, and seven) and their distribution using the integral performance indices: Integral Square Error (ISE) and Integral Absolute Error (IAE). The results demonstrate that the seven-function configuration achieved optimal values of 0.1155 and 7.365 ×10−4, for ISE and IAE respectively. In addition, this configuration achieved an energy efficiency of 99.7%, which is superior to the 98.9% efficiency obtained with three functions in the worst-case scenario. The optimal configuration was subjected to changes in irradiance and temperature to assess its dynamic response robustness. Additionally, an analysis of computational resource consumption reveals that the proposed hybrid controller requires a lower computational load for rule evaluation than three controllers from recent literature. These findings demonstrate the structural efficiency and optimization capability of the proposed system to maximize energy harvesting in photovoltaic panels at a low computational cost.

Abstract: The electro-hydraulic track tensioning system of a tracked vehicle directly affects track engagement stability, vibration response, and energy utilization efficiency under complex terrain and time-varying loads. Accurate and robust control is therefore of great engineering significance. This paper focuses on an electro-hydraulic tensioning system with a composite actuation structure consisting of a proportional main valve and two 2/2 on-off valves, and proposes a learning-enhanced nonlinear model predictive control method (Learning-enhanced Nonlinear Model Predictive Control, L-NMPC). Residual learning, adaptive weight/constraint scheduling, and execution-layer mode coordination are integrated into a unified predictive control framework. The study is carried out on a strongly coupled Simulink–AMESim–RecurDyn co-simulation model and an LF1352 prototype-vehicle test platform. Comparative evaluations are conducted under steady step-and-ramp tracking, random rough terrain, sudden steering/braking pulses, supply-pressure limitation, and parameter drift/sudden-change conditions. The evaluation indices include track-tension tracking error, peak overshoot, settling time, energy consumption, and stability under parameter mismatch. Compared with conventional nonlinear model predictive control (NMPC), the proposed L-NMPC reduces the root-mean-square error of track tension by 42%–58%, decreases peak overshoot by 30%–40%, shortens settling time by 25%–35%, and achieves a 12%–17% reduction in energy consumption at the simulation level. Under ±20% parameter perturbation, the fluctuation of track tension can be constrained within ±1.1 kN. The simulation and real-vehicle results remain consistent in terms of the dominant dynamic trends and performance ranking. This study provides a verifiable implementation path for model–data-fusion control of strongly coupled electro-hydraulic actuation systems and offers an engineering reference for intelligent, energy-efficient, and highly reliable control of tracked-vehicle chassis systems.

Abstract: This paper investigates the conflicting multiple constraints and safety challenges in humanoid robot teleoperation for nonprehensile transportation tasks. The robot's complex workspace and high degrees of freedom frequently conflict with highly dynamic task requirements, imposing stringent demands on coordinated motion. To address these issues, this paper proposes a Multiple-Constraint Safety-Critical Control Framework (MC-SCCF) featuring a hierarchical three-layer architecture. The top layer guarantees intrinsic safety against workspace boundaries using a continuously differentiable reachability surrogate model and an improved control barrier function (CBF)-based safe velocity filter for smooth deceleration. The middle layer maps user commands into pose-coupled reference trajectories to ensure task-level object safety, satisfying strict non-slip and non-toppling constraints. The bottom layer utilizes a quadratic programming (QP)-based inverse kinematics solver to achieve self-collision avoidance, coordinated motion, and optimal configuration while strictly enforcing joint and manipulability limits. Simulations and hardware experiments demonstrate that the MC-SCCF achieves real-time, high-precision reachability evaluation and successfully coordinates task dynamics with physical constraints, enhancing operational safety and the human-robot interaction experience.

Abstract: To address the issue of 2D laser-guided automated guided vehicles (AGVs) in industrial intelligent material handling scenarios being susceptible to interference from changes in lighting and complex obstacles, leading to abnormal positioning and mapping and frequent false stops, this paper designs a lightweight, multi-dimensional perception and anti-false-stop YOLOv8 anomaly recognition network, achieving accurate identification of various interferences in complex environments. An adaptive decision-making fault-tolerant control algorithm is proposed, introducing a temporal logic verification and dynamic threshold adjustment mechanism to achieve real-time dynamic switching of obstacle avoidance levels, ensuring efficient coordination between perception decision-making and control execution. An AGV anomaly detection sample set suitable for complex industrial scenarios is constructed, providing reliable data support for model optimization and accuracy evaluation. Finally, real-world deployment verification in a real electronics factory environment shows that this method reduces the vehicle false-stop rate and improves task handling efficiency. This research effectively solves the robust perception problem of AGVs in complex industrial environments and has significant engineering application value.

Abstract: Uniform control systems, that is multivariable control systems with identical transfer functions of separate channels and rigid cross-connections, are widespread in modern industry and technology, including such fields as mechatronics and robotics, electrical and aerospace engineering, chemical and power industry, and many others. In the paper, a sensitivity analysis of uniform control systems to small variations of parameters is given from the perspective of the characteristic transfer functions (CTFs) method. The formulas are derived determining the sensitivity functions of the CTFs and the canonical basis axes to small variations of parameters of uniform systems. The relations between the sensitivity functions of the open-loop and closed-loop uniform control systems are established. An illustrative example is considered concerning the sensitivity of control systems of multirotor un-manned aerial vehicles to small degradations of the motors’ effeciency.

Abstract: In the paper, a systematic treatment of sensitivity analysis of multivariable control systems from a perspective of the characteristic transfer functions (CTFs) method is given. The CTFs method (also called Characteristic Gain Loci method) allows one to associate with an N-dimensional multi-input multi-output (MIMO) system a set of N independent single-input single-output (SISO) characteristic systems and thereby to reduce the analysis and design of a MIMO system to analysis and design of N SISO systems. The formulas are derived determining the sensitivity functions of the CTFs and sensitivity vectors of the canonical basis axes to small variations of parameters of general type MIMO systems. The relations between the sensitivity functions of the open-loop and closed-loop MIMO systems are established. Two illustrative examples are considered. The first of them concerns the sensitivity of a two-dimensional not robust system with large degree of skewness of the canonical basis axes. In the second example, the sensitivity of the control system of a hexacopter (multirotor UAV with six rotors) to small degradations of the motors’ efficiency is analyzed.

Abstract: In the field of industrial robot vision, the accurate recognition and localization of transparent objects face multiple challenges. First, depth sensor data suffer from sparsity and non-uniform distribution. Even with high-end LiDAR, the obtained depth maps are generally sparse and severely noisy, especially around object boundaries. Most existing methods assume a fixed sparsity level, leading to significant performance degradation when the actual sparsity varies dynamically. Second, there is a cross-modal feature alignment issue between RGB and depth data. Simple channel stacking or addition neglects the modeling of feature correlations between the two modalities, resulting in insufficient information utilization. Furthermore, existing methods still lack the capability to model the multi-directional gradient variations of transparent objects under complex backgrounds. To address these issues, this paper proposes Attention-based Difference-enhanced Depth Fusion Network (ADDFNet), a depth completion network for transparent objects, which achieves synergistic improvements in accuracy and robustness through two key designs: MDAM and CMFR. To tackle the dynamic variation of sparsity and edge blurring, a Multi-directional Differential Attention Module (MDAM) is designed. It explicitly extracts multi-directional gradient information via multi-branch differential convolutions, enhancing the network’s robust perception of sparse edges. Within MDAM, a Detail Enhancement Differential sub-module (DEDM) and a Dynamic Convolution with Symmetry-enhanced Geometry Attention sub-module (DSCA) are introduced to adaptively adjust the focus regions under varying sparsity inputs. To address the insufficient cross-modal feature alignment, a Cross-Modal Feature Refinement (CMFR) module is introduced, which leverages RGB context to progressively guide and enhance depth features at the encoding stage, achieving finer cross-modal feature alignment. Evaluation results on the ClearPose and TransCG datasets demonstrate that ADDFNet outperforms comparison methods in terms of accuracy metrics.