Keywords: Learning
Abstract: We will consider policy optimization methods in reinforcement learning where the state space is countably infinite. The motivation arises from control problems in communication networks and matching markets. Specifically, we consider the popular Natural Policy Gradient (NPG) algorithm, which has been studied in the past only under the assumptions that the cost is bounded and the state space is finite, neither of which holds for the aforementioned control problems. Assuming a Lyapunov drift condition, which is naturally satisfied in some cases and can be satisfied in other cases at a small cost in performance, we design a state-dependent step-size rule which dramatically improves the performance of NPG for our intended applications. In addition to experimentally verifying the performance improvement, we also theoretically show that the iteration complexity of NPG can be made independent of the size of the state space. The key analytical tool we use is the connection between NPG stepsizes and the solution to Poisson’s equation. In particular, we provide policyindependent bounds on the solution to Poisson’s equation, which are then used to guide the choice of NPG stepsizes.

Keywords: Cooperative control
Abstract: In multi-agent systems—such as robotic fleets, autonomous vehicles, and industrial automation—efficient, conflict-free path planning is essential for safe, reliable operation. This talk explores innovative methods that enable multiple agents to navigate shared spaces without conflicts, balancing computational efficiency with real-world constraints. Practical approaches will be covered, including optimization techniques, heuristic algorithms, and decentralized frameworks, with a focus on scalability and adaptability to dynamic obstacles. Special emphasis will be placed on handling complex, high-density environments where collision avoidance is critical. Attendees will gain insights into strategies that enhance pathfinding efficiency, minimize conflicts, and provide robust solutions for complex, real-world applications across diverse multi-agent scenarios.

Keywords: Cooperative control
Abstract: Cohesion in networks during transitions from one consensus value to another, i.e., the ability of agents to respond in a similar manner during the transition, can be as important as achieving the new consensus value. Existing decentralized network control strategies mainly concern with the convergence speed to the final consensus value. However, even with increased convergence speed, the level of cohesion loss during transitions can be large. This loss of cohesion during transition (and tracking of varying consensus values) can be alleviated using a recently developed delayed self-reinforcement (DSR) approach. However, the current DSR-based approach assumes ideal conditions with agents having instant access to neighbor information – without network delays arising during sensing or communication between neighbors, as well as computation of control actions of each agent, which can cause instability. In this talk, I will present my work on addressing the issue of cohesive transitions in networks with delays. Additionally, I will discuss the results of cohesion control in formation control of multi-agent systems, multi-robot flexible object transport and inter-vehicle spacing control in connected vehicles.

Keywords: Stability of linear systems
Abstract: With the advances in sensing technology, computing capabilities, and storage infrastructure there has been a surge in using data for an increasing number of applications. However, the vast literature on data based analysis and control is mostly focussed on dynamical systems that evolve over one independent variable, namely, time. In this talk, we shall discuss about data-based conditions for checking stability of a special class of systems having more than one independent variable. It is known that because of the lack of natural ordering of the domain, the problem of defining and studying stability of systems having n independent variables (nD systems) is not straightforward. We focus on the following two notions of stability: (i) stability with respect to a given direction, and more generally, (ii) conic stability, that is, stability with respect to subsets of the domain having the structure of a cone, and provide data-based tests for checking them.

Keywords: Nonholonomic systems
Abstract: Moving platforms such as ships with firing turrets, mobile air-defense systems, surveillance aircraft, and military vehicles with directed sensors are all examples of moving pursuers equipped with rotating platforms. The platforms can be a turret, a directed energy weapon, or a gimbaled camera and sensor. This talk will explore time-optimal and energy-optimal motion planning strategies for such vehicles. We will begin by introducing a novel vehicle model, called the Dubins-Laser system. Following this, we will discuss how first-order necessary optimality conditions can be leveraged to reduce the trajectory search space to a finite set of candidate solutions. Finally, numerical simulations will be presented to demonstrate the practical applications of the proposed analytical results.

Keywords: Quantum control, Emerging control theory, Emerging control applications
Abstract: In this paper, we present a feedback-based framework for controlling the evolution of quantum spin states of nuclei within an ensemble. This approach allows for the selective excitation of nuclear spin states within specific pass bands of Larmor frequencies while leaving the nuclei outside the pass bands undisturbed. The control input sequence is designed in simulation and implemented using radio frequency pulses in the real system.

Keywords: PID control, Automotive, Simulation
Abstract: This paper addresses vehicle platooning on curved path or curved platooning and presents an algorithmic solution for platooning homogeneous position-controlled vehicles (PCVs) under two communication topologies: leader-following (LF) and predecessor-following (PF). Most existing works present curved platooning with constant speed. On the other hand, this work demonstrates curved platooning with varying speeds. In this paper, all the follower vehicles receive the lead vehicle's communicated actual trajectory, i.e., its actual position and orientation, for platooning with the LF topology. In LF topology, a follower vehicle remains unaware of its preceding vehicle's maneuver and PF topology overcomes this limitation. Therefore, this paper investigates platooning with PF topology using the reference as well as the actual trajectory of the preceding vehicle in two separate approaches. In all the approaches, a follower vehicle creates a virtual trail of the lead or the preceding vehicle by storing the communicated information from the respective vehicles. An algorithm on the follower vehicle uses the virtual trail and finds the follower vehicle's reference position and orientation on that trail, maintaining a constant inter-vehicle distance. This work compares the applicability of the algorithmic solutions with the LF and PF topologies and presents the simulation results.

Keywords: Robust adaptive control, Mechanical systems/robotics, Biomedical
Abstract: Micro/nanorobots have the potential to revolutionize the medical field by performing tasks such as targeted drug delivery and minimally invasive treatments. For these tasks, path tracking is a critical objective, achievable through closed-loop feedback controllers. However, traditional control methods may be insufficient for accurate tracking, as micro/nanorobots operate in low Reynolds number fluids inside the body, facing various uncertainties and unknown disturbances. To address these challenges, this paper presents a robust adaptive control method for the precise path tracking of microrobots. A time-delayed controller is designed, which utilizes past information to estimate system dynamics uncertainties. The controller parameters, including controller gains and artificial delay values, are obtained using a Lyapunov-Razumikhin-based analysis. This controller design is validated through MATLAB-based simulations and a developed in vitro experimental setup. With the proposed approach, the microrobot is able to track reference trajectories. These results will contribute to advancing robotics applications in the healthcare domain.

Keywords: Mechanical systems/robotics, Control applications, Mechatronics
Abstract: The constrained spaces in narrow aisles of airplanes and manufacturing facilities limit the use of mobile robots. Utilizing Mecanum wheels allows mobile robots to move smoothly in any direction without needing to turn, enhancing maneuverability and efficiency in constrained environments. However, the sizeable Mecanum wheel and the individual motor assembly required for independent control restrict its use in mobile robot designs for maneuvering in narrow spaces. This work proposes an asymmetric design to overcome the challenge of using Mecanum wheels in small-size robots and analyses the kinematic requirements due to asymmetric design. Numerical simulation and experimental tests confirm the platform’s ability to navigate tight spaces, make sharp turns, and carry various payloads, significantly improving operational efficiency. The proposed asymmetric mobile robot offers a practical solution for enhancing the performance of automated guided vehicles (AGVs) and robotic systems in space-limited settings.

Keywords: Control applications, Robust adaptive control, Uncertain systems
Abstract: The grasp control of a robotic gripper poses unique challenges to roboticists owing to the wide horizon of applications associated with robotic end-effectors. The typical control problem requires a gripper to securely grasp an unfamiliar object in an uncertain environment. The complexity of the problem gets compounded with fragile or deformable objects included as candidate grasp-subjects. An intelligent grasp algorithm regulates the grip force using sensor feedback to ensure a secure grasp even in a highly uncertain scenario. This paper attempts to use a slip detection-based feedback scheme that employs a simple and intuitive gain-scheduling-based control law to effectively adapt the grasp force and ensure a secure grasp even under extreme parameter perturbations.

Keywords: Feedback linearization, Nonlinear systems, Control applications
Abstract: This paper presents the development and implementation of a real-time state-to-state feedback linearization controller for an inertia wheel pendulum. Based on the nonlinear model of the plant, a mathematical proof of existence of such controller is first given, then a non-trivial local diffeomorphism is found that puts the system in the normal form, and finally the latter is employed to design a nonlinear control law that performs state-to-state exact feedback linearization to maintain the inertia wheel pendulum in the vertical upward position. Details guaranteeing reproducibility of results are provided, both for simulation and real-time implementation.

Keywords: Optimization, Power systems, Nonlinear systems
Abstract: Insufficient instrumentation and uncertainty in network information hinder the task of voltage estimation in electrical distribution networks. Considering this challenge, in this work, we propose a model-based algorithm that estimates the state (voltage) and the parameters (network impedance) of a single-phased radial power distribution network. We model this network leveraging the recently introduced lossy distribution flow equation. We formulate the estimation problem in the spirit of a constrained optimization framework, which exploits available measurements of network voltage and bus power injection data as ground truth and generates fast voltage and impedance estimates. We theoretically verify the convergence of voltage and impedance estimation errors embracing results from control Lyapunov function and perturbation theory. We simulate our algorithm on the standard IEEE 33-bus distribution system and observe that it provides improvements in both voltage and impedance estimates, in comparison with the traditional power flow models that ignore network losses

Keywords: Uncertain systems, Nonlinear systems, Stochastic systems
Abstract: In this paper, the controllability property for Caputo fractional delay Partial Random Differential Equations (PRDE) in Banach space are investigated. By utilizing the principles from fractional calculus, measures of non-compactness, and fixed point theorems within a stochastic framework, the existence of random mild solutions are established. The Mddot{o}nch fixed point theorem is then employed to examine the controllability of the proposed system. An example is given to validate the theoretical results.

Keywords: Stability of nonlinear systems, Power systems, LMIs
Abstract: This paper investigates the switched load frequency control techniques for nonlinear power systems together with energy storage systems (ESS) subject to actuator faults. In this regard, the faults of actuator are assumed to occur randomly and the fault rates are characterized by stochastic variables that satisfy certain probability conditions. Next, a Takagi-Sugeno fuzzy model is employed to describe the nonlinear characteristics in turbine and governor dynamics of power systems with ESS. The reliable proportional-integral (PI)-type load frequency control is proposed by utilizing the switching approach and Lyapunov-Krasovskii functional with the membership function information, which ensures the asymptotic stability of power system with ESS in the mean-square sense via H_{infty} performance index. At last, the numerical examples are presented to illustrate the superiority of the proposed method.

Keywords: Constrained control, Nonlinear systems
Abstract: The paper addresses the problem of controller synthesis for control-affine nonlinear systems to meet reach-avoid-stay specifications. Specifically, the goal of the research is to obtain a closed-form control law ensuring that the trajectories of the nonlinear system reach a target set while avoiding all unsafe regions and adhering to the state-space constraints. To tackle this problem, we leverage the concept of the funnel-based control approach. Given an arbitrary unsafe region, we introduce a circumvent function that guarantees the system trajectory to steer clear of that region. Subsequently, an adaptive funnel framework is proposed based on the target, followed by the construction of a closed-form controller using the established funnel function, enforcing the reach-avoid-stay specifications. To demonstrate the efficacy of the proposed funnel-based control approach, a series of simulation experiments have been carried out.

Keywords: Estimation, Large scale systems, Observers for linear systems
Abstract: This paper presents the design of control for a linear time-invariant system with three different time scales using multirate state feedback (MRSF). The term ``multirate sampling’’ in this context refers to the act of sampling the slow, fast, and fastest states at varying time intervals. Initially, the original continuous three-time-scale system is transformed into a block triangular form. Subsequently, it is discretized in a step-by-step manner, and feedback controls are simultaneously devised at different sampling rates to create a unified MRSF control. It is established that this derived MRSF control stabilizes the entire system. Additionally, a sequential three-stage observer is recommended for estimating the transformed slow, fast, and fastest states, which is crucial for implementing MRSF. Lastly, the design is illustrated through a numerical example.

Keywords: Control of networks, Large scale systems
Abstract: This work investigates the tracking of dynamic trajectories using consensus-based decentralized algorithms. Dynamic average consensus (DAC) algorithms seek to improve tracking of the average signal by increasing the convergence rate, e.g., using accelerated versions of standard DAC algorithms. The main contribution of this work is to show that improving the convergence rate doesn’t necessarily ensure improved tracking of the average signal, especially during rapid changes in the measured signals. Additionally, this work also presents a DAC-based delayed self reinforcement (DSR) approach to improve the tracking of the average signal rather than the convergence rate. The proposed DSR-DAC algorithm, which approximates the ideal centralized DAC, improves tracking of the average signal by only using already-available (current and past), local information from the network and sensed reference signals. Numerical simulation results for a multi-mobile agent tracking problem demonstrate substantial improvement with 91% and 84% reduction in tracking error using the proposed DSR-DAC approach when compared to the standard and accelerated DAC algorithms, respectively.

Keywords: Decentralized control, Large scale systems, Autonomous systems
Abstract: This paper presents a battery-aware hierarchical task allocation (B-HTA) algorithm, which combines a higher-level centralized and lower-level decentralized task allocation layers, for multi-robot systems in warehouse operations. Fully centralized control architectures can provide more equitable task distribution among robots, but also rely extensively on communication which can be costly, hard to scale with number of robots and prone to uncertainties. On the other hand, a decentralized control framework is scalable and robust, however, it cannot achieve similar balanced task allocation as centralized control due to lack of global information. Balanced task allocation for warehouse operations is essential for effective use of the robot resources and minimize time losses in recharging. This paper proposes a blended approach of a hierarchical structure, which is scalable and requires minimal communication, consisting of two layers, i) a centralized layer that does zone wise task allotment, and, ii) a decentralized layer based on market auction algorithm that allots tasks to robots in their respective zones. The decentralized layer of this algorithm takes into consideration the battery level of the robot while allotting it a task, which leads to more balanced task allocation among the robots and fewer recharging instances, as shown through simulation results compared to a completely centralized approach and also outperforms other decentralized market auction based approaches.

Keywords: Large scale systems, Stability of linear systems, Control of networks
Abstract: Urban transportation networks often encounter delays that cascade as secondary disruptions throughout the system. While delay prediction using historical or simulated data is well-established, there remains a gap in accurately forecasting secondary delays propagated through interconnected services within the network without relying on periodicity assumptions. In this paper, we introduce a secondary delay propagation algorithm for suburban operations within a division of Indian Railways. Under the assumption of a static timetable and fixed network topology, this algorithm effectively predicts future delays based on initial primary delays. Additionally, we introduce the concept of systemic stress and examine its redistribution in the presence of secondary delays. We demonstrate resilience as the ratio of recovered services to affected services and rapidity as the time required for the network to return to its regular schedule. Finally, we conclude with the framework's potential to predict dynamic freight corridors in real-time, enabling railway companies to operate unscheduled, profitable freight services without disrupting passenger operations.

Keywords: Linear systems, Optimization, Large scale systems
Abstract: We consider the control of discrete-time linear dynamical systems using sparse inputs where we limit the number of active actuators at every time step. We develop an algorithm for determining a sparse actuator schedule that ensures the existence of a sparse control input sequence, following the schedule, that takes the system from any given initial state to any desired final state. Since such an actuator schedule is not unique, we look for a schedule that minimizes the energy of sparse inputs. For this, we optimize the trace of the inverse of the resulting controllability Gramian, which is an approximate measure of the average energy of the inputs. We present a greedy algorithm along with its theoretical guarantees. Finally, we empirically show that our greedy algorithm ensures the controllability of the linear system with a small number of active actuators per time step without a significant average energy expenditure compared to the fully actuated system.

Keywords: Stability of linear systems, LMIs, Fault-tolerant systems
Abstract: This paper explores the problem of stabilization and the H∞ performance of a 2-degree-of-freedom (2-DoF) helicopter system (HS) subject to bounded external perturbation via sampled-data control. Meanwhile, the random variable, which follows a Bernoulli distribution, is established to characterize the stochastic actuator faults. Then, a looped Lyapunov-Krasovskii functional (LKF) is constructed with complete information throughout the sampling interval. Based on the looped LKF and robust control, new stability criteria for the closed-loop HS are obtained in the form of linear matrix inequalities (LMIs). Thus, the proposed sampled-data control for the given HS model ensures the asymptotic stability and satisfies the H∞ performance index. Finally, the effectiveness of the proposed theoretical concepts are validated through numerical results, in achieving robust stability and diminishing perturbation attenuation levels.

Keywords: Behavioral systems, Emerging control applications, Linear systems
Abstract: The notion of dissipativity is immensely important for system analysis and controller design. In this paper, we formulate data-driven conditions for verifying dissipativity properties of a discrete linear time invariant dynamical system. Most data-driven conditions in the literature guarantee dissipativity only for a finite interval of time. Also, most data-driven methods that guarantee dissipativity consider what is referred to as the (Q,S,R) supply rate. We derive our data-driven dissipativity conditions for general supply rates which involve time differences of measured data. Moreover, our data-driven conditions guarantee dissipativity for infinite time horizon. We make use of the behavioral framework of dynamical systems and the notion of quadratic difference forms (QdFs) to formulate our data-driven dissipativity conditions as they help in generalizing the notion of dissipativity to general supply rates. Moreover, formulation of our data-driven dissipativity conditions does not require assuming persistence of excitation of measured data, which is a necessary assumption in formulation of data-driven dissipativity conditions proposed in the literature.

Keywords: Predictive control for linear systems, Optimization algorithms, Hybrid systems
Abstract: This paper presents a Discrete-Time Model Pre- dictive Controller (MPC) for humanoid walking with online footstep adjustment. The proposed controller utilizes a hier- archical control approach. The high-level controller uses a low-dimensional Linear Inverted Pendulum Model (LIPM) to determine desired foot placement and Center of Mass (CoM) motion, to prevent falls while maintaining the desired velocity. A Task Space Controller (TSC) then tracks the desired motion obtained from the high-level controller, exploiting the whole- body dynamics of the humanoid. Our approach differs from existing MPC methods for walking pattern generation by not relying on a predefined foot-plan or a reference center of pressure (CoP) trajectory. The overall approach is tested in simulation on a torque-controlled Humanoid Robot. Results show that proposed control approach generates stable walking and prevents fall against push disturbances.

Keywords: LMIs, Decentralized control, Stability of linear systems
Abstract: This paper investigates the stochastic stability and stabilization issues of a linear interconnected Markovian jump system (LIMJS) with time-varying delay in the interconnected term. To address this, a mode-dependent sampled-data control is designed, and a new looped-Lyapunov functional (LLF) is constructed, utilizing the information of the sampling instants and time delay. The linear matrix inequality-based, delay-dependent sufficient conditions (DDSCs) are derived with the help of the new LLF, which ensures the stochastic stability of LIMJS with mathscr{H}_{infty} performance level. Finally, a numerical example and its simulation results are given to illustrate the effectiveness of the proposed DDSCs.

Keywords: Delay systems, Linear systems, LMIs
Abstract: This paper focuses on designing the anti-windup (AW) compensator for continuous time-delayed systems with actuator saturation (AS) using the guaranteed cost control (GCC) approach. The study specifically considers linear time-delayed systems with AS and takes into account both the AW design and stability analysis of the system. To compute the AW compensator gain, a linear matrix inequality (LMI)-based study is used. Some suitable examples are used as a demonstration of the approach presented.

Keywords: Mechanical systems/robotics
Abstract: In cable driven parallel robots (CDPRs), the payload is suspended using a network of cables whose length can be controlled to maneuver the payload within the workspace. Compared to rigid link robots, CDPRs provide better maneuverability due to the flexibility of the cables and consume lesser power due to the high strength-to-weight ratio of the cables. However, amongst other things, the flexibility of the cables and the fact that they can only pull (and not push) render the dynamics of CDPRs complex. Hence advanced modelling paradigms and control algorithms must be developed to fully utilize the potential of CDPRs. Furthermore, given the complex dynamics of CDPRs, the models and control algorithms proposed for them must be validated on experimental setups to ascertain their efficacy in practice. We have recently developed an elaborate experimental setup for a CDPR with three cables and validated elementary open-loop motion planning algorithms on it. In this paper, we describe several aspects of the design and fabrication of our setup, including component selection and assembly, and present our experimental results. Our setup can reproduce complex phenomenon such as the transverse vibration of the cables seen in large CDPRs and will in the future be used to model and control such phenomenon and also to validate more sophisticated motion planning algorithms.

Keywords: Networked control systems, Sensor fusion, Estimation
Abstract: This experimental work introduces a real-time hardware-in-the-loop framework using software-defined radios, serving as a testbed for implementing networked control systems with real hardware. This work also considers an artificial channel modeler to emulate time varying channel uncertainties. The overall system enables the remote distribution of modules such as sensors, controllers, and estimators, mimicking real-world setups. One test case features a local plant with a sensor and a remote node containing both an estimator and a controller. This framework bridges the gap between theoretical research and practical implementation, providing a robust platform for pre-deployment testing.

Keywords: Aerospace
Abstract: Abstract - The work reported in this paper presents experimental research on sine sweep frequency-controlled Lead Zirconate Titanate (PZT) piezoceramic patches to de-ice a typical leading edge of an aircraft wing. The PZTs are excited in d31 mode to induce vibration so that induced out-of-plane shear stresses to shear off the deposited ice. The modular concept of leading-edge curved profile in the aircraft wings between the cross ribs can be effectively utilized to implement this electromechanical excitation for de-icing. Experimental vibratory deicing is effective at resonance frequencies first at twist (torsion modes) and subsequently at bending modes. This process ensures higher out-of-plane shear stresses that are more than the adhesive shear strength of ice on an aluminum surface typically 1.5 MPa. Sine sweep frequency-controlled excitation method is found to be the best since it excites the bending and twist modes and takes care of the time variant aircraft wing leading edge with ice. The ratio of area of PZT to total wing area is only 1.8%. The power consumed by this electromechanical excitation is of the order of 0.35 kW/m2 in comparison to 10 kW/m2 deicing by Electro-Thermal method. This method is also environment compatible in contrast to the conventional chemical method that contaminates the ground water. The results indicate a possibility of potential applications of the PZTs in deicing as they can be used both as ice thickness detecting sensors as well as actuators for deicing.

Keywords: Control applications, LMIs, Linear systems
Abstract: We consider the problem of designing and implementing feedback controllers, based on data, for a lab-scale rotary flexible link system. We first consider the data-driven stabilization problem and show that while the system can be stabilized using a data-driven state-feedback controller, the reference tracking with this controller is not satisfactory. For tracking a rotary servo angle reference, we then design a data-driven reference tracking controller. We provide necessary and sufficient design conditions, in terms of data-based linear matrix inequalities. The controllers are also implemented in real-time.

Keywords: Mechatronics, Mechanical systems/robotics, Manufacturing systems
Abstract: Box-type solar cookers are portable and easy-to-transport renewable cooking devices that can replace polluting fuels in small households and remote areas, providing health benefits for women while also benefiting the environment. However, a lengthy cooking time hinders widespread adoption. While researchers have investigated different approaches to shorten cooking time including passive tracking mechanisms, active tracking has been mostly overlooked. The aim is to introduce a new two-axis tracking mechanism for box-type solar cookers (BTSC) and design a control strategy that automatically detects and follows the sun's position at any location. The effectiveness of tracking is evaluated based on the optimal orientation ratio, the first figure of merit (F_{1}), the second figure of merit (F_{2}), and cooking power. The suitable tracking intervals for maximizing energy output are determined by balancing actuator energy consumption with discrete tracking steps. These statistics demonstrate that active cooker tracking is feasible and yields well.

Keywords: Kalman filtering, Sensor fusion, Machine learning
Abstract: A rapid technological advance in data science and push for digitalization in the process industry resulted in the increasing popularity of data-driven-based predictive solutions/decision-making in the process industry in recent years. Data-driven soft sensors or predictive models, which are easier to develop and implement, are proven to be alternative solutions for costly hardware analysers. The accuracy of the data-driven soft sensors or machine learning-based predictive models highly depends on the data quality. Soft Sensor design for industrial processes often involves handling measurements at different rates and delays. Slow-rate, irregular, and delayed measurements are generally more accurate but suffer from delays due to analysis time, such as laboratory testing. In contrast, fast-rate measurements, such as flow rate/temperature measurements, are available at regular intervals without delay but tend to be less accurate. Further, since these predictive models are developed based on past historical data, developed under normal operating conditions, the predictive performance of these data-driven soft sensors/machine learning-based predictive models deteriorates in the presence of unmodelled faults in the processes. Unknown faults, such as catalyst deactivation, leaks in the process tanks, etc. need to be detected or estimated for accurate estimation of the states of the system. To improve the prediction performance of data-driven soft sensors/machine learning-based predictive models in the presence of noisy data, this work demonstrates the implementation of a machine learning-based state estimator on experimental hybrid three tank system. Here, a machine learning-based state estimator is developed by integration of dynamic Deep Neural Networks with the Bayesian State Estimators, such as linear or nonlinear Kalman filters. Analysis of the result indicates that the proposed machine learning based state estimator accurately tracks the level measurements and addresses the noisy data issue.

Keywords: Control education, PID control, Control applications
Abstract: Maintaining a precise liquid level is crucial for safety and efficiency in many industries. Various control algorithms can be used to keep the same, such as Model predictive control (MPC), Proportional-Integral-Derivative (PID), etc. These controllers use sensor feedback to monitor the liquid level and adjust valves or pumps accordingly. In this paper, we describe an experimental test bed equipped with a low-cost level sensor and an in-house developed valve actuation system, which can be combined with any commercially available quarter-turn valve to conduct an educational experiment to help students understand the theory and functional logic behind the PID based liquid level controller.

Keywords: Control of networks, Optimal control, Discrete event systems
Abstract: In the localities where there is a limitation on water availability, water supply will be available only for a certain period in a day and in rural areas water collection will be possible only during these limited periods. In this type of networks, inefficient operational methods leads to inequity in water supply. This paper uses data driven method to schedule the supply of available water. A scheduling problem is formulated and solved using the flow measurements from a scaled real time network. A heuristic is developed using Epsilon greedy algorithm to limit the number of state measurements. This approach was tested on a lab scale Water Distribution Networks (WDN) and optimal scheduling was found with minimal measurements.

Keywords: Control applications, Mechatronics, Control education
Abstract: This paper introduces a novel soft robotic finger (SRF) fabricated from Polydimethylsiloxane (PDMS), designed to perform controlled bending and yaw movements. Various experiments are conducted to demonstrate the SRF’s performance in real-time. The bending movement is driven by an internal actuation channel, where the yaw movement is performed by a dual-motor system. Embedded sensors continuously monitor and adjust bending and yaw movement, improving control during gripping, lifting, and turning objects.This work advances to design a SRF at low cost by enhancing continuous control operation with material flexibility.

Keywords: Agents-based systems, Automotive
Abstract: We present comparative methods for detecting lanes and extracting lane points in scaled-down vehicles with Edge computing devices. The approach involves constructing a custom-made track using coloured tapes to define lane boundaries and extract lane boundary information from images captured using a camera mounted at the front of a scaled-down electric vehicle. The edge detection algorithms identifies the most significant edges within an image or scene. These detected edges are then used to determine meaningful lines and boundaries; in our case, we isolate the lane boundaries. The study evaluates three edge detection techniques — Sobel filter, Canny edge detector, and Gabor filter, to generate an edge map from the camera image frames. To further refine the detection process, we apply colour masks and isolate the relevant lane regions more effectively. The system is implemented on a scaled-down electric vehicle platform, named DEFT, capable of autonomous motion. The vehicle is equipped with a ZED Mini stereo camera and interfaced with a Jetson Nano via USB. We provide target angular velocities for the driven wheels and utilize a PID controller to track the desired RPM. We steer the vehicle via teleoperation using a gamepad and run the vehicle over the developed track with coloured lane boundaries. Finally, we provide a performance comparison (quality, response time, and accuracy) for the filters at different vehicle speed and highlight the best choice for lane detection in scaled-down vehicles with Edge computing devices.

Keywords: Quantum control
Abstract: This lecture will present a collection of results in the area of linear quantum control engineering. It will discuss models for quantum systems using the Heisenberg pictures of quantum mechanics to describe continuous linear quantum systems. It will discuss closed loop approaches to quantum control involving coherent quantum feedback control in which the controller is also a quantum system. It will discuss quantum H - infinity control, quantum LQG control, coherent quantum observers and coherent quantum state estimation. The lecture will also cover the quantum Kalman decomposition as well as recent results on quantum risk sensitive control and networks of quantum linear systems as quantum memories. Applications in the areas of quantum optics and quantum electromechanical systems will also be presented.

Keywords: Optimization
Abstract: Optimization algorithms are the driving force behind many machine learning formulations, and recent progress in this field has been greatly aided by a better understanding of these algorithms and their implementation for specific machine-learning problems. However, the field lacks a generalized tool to design and analyze optimization algorithms, making it non-intuitive for practitioners to develop their accelerated variants. The study of gradient flow and its connection to dynamical systems has a long history in mathematics. This talk aims to further explore this relationship from a continuous-time perspective. In particular, we introduce a generalized framework for designing accelerated optimization algorithms based on recent advances in the fixed-time stability theory of continuous-time dynamical systems. This framework allows for the strongest possible convergence guarantees and easily extends to a subclass of non-convex functions. This talk is part of a larger effort to connect the fields of dynamical systems and optimization and to advance our understanding of the underlying mathematical principles and their applications.

Keywords: Machine learning
Abstract: Software-defined vehicles (SDVs) are redefining automotive technology. They integrate modular architectures like SOFAEE and Autoware with emerging zonal designs to enhance scalability and control. These architectures optimize resource utilization and decentralize subsystems, addressing fault tolerance and real-time processing challenges. The integration of AI/ML introduces new dimensions for perception, decision-making, and control, but also demands robust strategies for real-time optimization and cybersecurity. With SDVs increasingly connected, safeguarding against cyber threats while ensuring functional safety has become critical. This presentation explores control strategies, system-level challenges, and collaborative solutions necessary for advancing SDV deployment. It highlights the importance of adaptive control frameworks, open architectures, and standardized protocols to address these complexities. By bridging control theory with emerging technologies, this discussion provides actionable insights for researchers and practitioners to accelerate innovation in autonomous mobility.

Keywords: Cooperative control, Autonomous systems, Agents-based systems
Abstract: This paper addresses the formation control problem in multi-agent systems with a central flock leader. Within this framework, each agent of the flock exchanges its position and velocity information with their neighbors. The navigation command for the entire group is determined by the leader's configuration, while followers adopt a lattice formation based on inter-agent communication, which is limited to the sensing range of an agent. The proposed algorithm enables the leader bot to position itself while other agents maintain the desired formation. The flock can also move with a variable velocity while being in formation. The effectiveness of the proposed leader-driven flocking algorithm is demonstrated using MATLAB and Gazebo simulation environment. Finally, the real-time implementation of the proposed algorithm on the actual JetBots demonstrates its practical applicability.

Keywords: Multivehicle systems, Cooperative control, Control applications
Abstract: This paper explores Morse Potential based Attraction-Repulsion (MPAR) methodology in a modified cyclic pursuit framework to achieve swarm aggregation leading to circumnavigation behaviour. However, MPAR, being a nonlinear range-dependent function, displays complex drifting behavior and arbitrary shape of the swarm. This complexity is sensitive to the initial conditions of the agents and appears to be analytically intractable, thus making it unusable in many applications. We address this issue by making one of the agents stationary and modifying the cyclic network. It is shown that circumnavigation is achieved with the target information available to only one agent. Furthermore, based on the initial velocity perturbation of a follower agent, some non-intuitive aggregation patterns emerge. We use a LOS-based analysis framework to explain these patterns, which can be used for location-sensitive applications and for search and exploration.

Keywords: Nonholonomic systems, Decentralized control, PID control
Abstract: UAV swarms have gained prominence due to their reliability, agility, and cooperative capabilities. This paper addresses the challenges of devising algorithms for fixed-wing UAVs posed by non-holonomic kinematics and actuator constraints. While existing works explored optimal control, artificial potential fields, and biologically inspired methods, these often overlook 3-dimensional non-holonomic constraints and actuator saturation. This work contributes to a decentralized 3D fixed-wing flocking algorithm that ensures inter-agent collision avoidance and effective flock formation. Moreover, using a leader-follower model, the agents track a virtual leader while communicating only with neighbors, keeping the motion constraints of fixed-wings. We establish safe initial conditions to guarantee collision avoidance. Numerical simulations validate the proposed algorithm's robustness, demonstrating its effectiveness in solving the flocking problem for non-holonomic fixed-wing agents and advancing UAV swarm coordination.

Keywords: Cooperative control, Autonomous systems
Abstract: This paper proposes a distributed algorithm to drive all agents in a heterogeneous multi-agent system (MAS) to the geometric mean of their initial conditions. It is defined as the GM-consensus problem. The proposed GM-consensus algorithm has a distributed observer for estimating the geometric mean of agents' initial states. Its later part is a local controller for tracking these observed states. This structure facilitate the scaling of the algorithm for the agents having higher-order model. The numerical example exemplifies the efficiency of the proposed distributed GM-consensus algorithm.

Keywords: Cooperative control, Autonomous systems
Abstract: This paper proposes a containment control law for a heterogeneous multi-agent system (MAS) by using the binary relative position measurements of the agents and the targets. Unlike the existing methods, this approach relaxes the necessity for exact position measurements as feedback to the agents, thereby the control protocol can be executed with less accurate sensors. To the best of authors' knowledge, this is the first attempt to design the containment law for a heterogeneous MAS based on the binary relative position measurements. Using non-smooth Lyapunov analysis, we establish the convergence of the proposed containment law under a detailed-balanced directed network with any positive control gain. The numerical simulations substantiate the efficacy of the proposed binary measurement-based containment law.

Keywords: Stochastic systems, Optimal control, Cooperative control
Abstract: This research focuses on identifying the necessary and sufficient conditions for achieving Pareto optimality in delayed cooperative differential games that are governed by mean-field stochastic differential equations over a finite horizon. We express the game problem as a set of constrained delayed mean-field stochastic optimal control problems. Through the application of the maximum principle for optimal control in mean-field stochastic delay differential systems, we derive the necessary condition for a Pareto efficient strategy. Additionally, we provide the sufficient condition under the convexity assumptions on certain functions. Finally, a linear quadratic optimal control problem is considered as an application of the theoretical findings.

Keywords: Aerospace, Nonholonomic systems, Networked control systems
Abstract: This paper presents a distributed guidance law for the simultaneous interception of a stationary target using a group of `n' heterogeneous pursuers. Under the proposed law, simultaneous interception of the target is guaranteed to happen for any arbitrary initial positions or heading angles of the pursuers. A cyclic communication topology is used in this paper, where every pursuer interacts only with its neighbour. We also determine the necessary conditions to impose a feasible desired impact time. The theoretical results have been illustrated through numerical simulations.

Keywords: Networked control systems, Cooperative control, Output feedback
Abstract: This paper addresses the output consensus problem for networked mobile robots by using negative-imaginary (NI) systems theory. It is shown that there exists a nonlinear state feedback control law for each wheeled mobile robot (WMR) which results in satisfying the output strictly negative-imaginary (OSNI) property of the closed-loop system. Nonlinear NI controllers are placed in positive feedback with the rendered OSNI systems through a fixed network topology. It is shown that the nonlinear NI/OSNI properties are retained in the networked configuration and the output consensus result is proved. The proposed result is robust against NI uncertainties and is applicable to multi-robot systems without considering any linearization assumption. Simulation results are presented to validate the efficacy of the proposed control scheme.

Keywords: Control of networks, Networked control systems, Decentralized control
Abstract: This paper investigates the problem of state consensus in multi-agent systems (MASs) under periodic denial-of-service (DoS) attacks affecting the inter-agent communication network. We introduce the so-called deviation matrix to quantify the impact of such DoS attacks on the consensus point of MAS with that of the average consensus point under no attack. It is demonstrated that the agents achieve asymptotic consensus at a perturbed location under DoS attacks of a finite period where the deviation between the average consensus and the perturbed consensus points is established analytically and found to be dependent on the DoS attack frequency and duration. Additionally, with an infinite attack period and communication link between any set of agents compromised, the MAS behaves as a decoupled set of agents, each achieving equilibrium at the origin. Illustrative examples and simulation results are provided to validate the theoretical developments in the paper.

Keywords: Control of networks, Networked control systems
Abstract: In this paper, the Laplacian network dynamics is considered, in which every node follows the classical nearest-neighhbour rule. One of the nodes, referred as the control node, is allowed to receive the external control input. To quantify the control energy required for the state transfer, the average controllability centrality (ACC) is prevalently used in the literature. The ACC of a node is defined as the trace of the Controllability Grammian (CG) matrix of the network dynamics when that node is acting as the control node. In this paper, the effect of network topology and control horizon, on the ACC is investigated. Subsequently, it is shown that under either of the following conditions, the ACC’s of all nodes are almost equal: i) Networks with high edge density ii) Too short and too long control horizon As a result, it is concluded that under these conditions, the ACC is not an efficient metric for selecting the energy-efficient control node. Finally, the proposed results are demonstrated through simulations on randomly generated graphs.

Keywords: Networked control systems, Control of communication networks, Linear systems
Abstract: This paper deals with the problem of stabilization of a multi-loop networked control system over a shared communication network. We co-design a polynomial control law and a communication scheduling strategy, which is based on the earliest deadline first (EDF) algorithm. The proposed polynomial control strategy allows for generating a time-varying control input to the plant of each control subsystem, even between two successive communication instants, by transmitting limited information over the communication network and by using limited computational resources at the actuator. We provide a sufficient condition that ensures that the proposed communication scheduling strategy guarantees global asymptotic stability of the origin of each control subsystem. We illustrate our results through numerical examples.

Keywords: Networked control systems, Linear systems, Stability of linear systems
Abstract: The objective of this paper is to explore the property of eventual exponential positivity (EEP) in complex matrices. We show that the property of EEP can be guaranteed for the real part of the matrix exponential for a certain class of complex matrices. Thereafter, we present the relation between the spectral properties of a Laplacian matrix of an unsigned digraph with complex edge-weights and the property of real EEP. Applying these results to a Laplacian flow system, we show that the stability of such a system is ensured when the negated Laplacian is real EEP for the given network. Numerical examples along with a counterexample are presented to demonstrate the results.

Keywords: Identification, Estimation, Robust adaptive control
Abstract: Control frameworks for legged robots often rely on accurate dynamic models. However, these models often proves to be inaccurate due to factors such as mechanical wear and tear, and unforeseen changes such as the addition of extra payloads during deployment. Significant deviations in the dynamics can severely impact the controller’s performance. Our goal is to enhance the controller’s model in real-time during deployment using onboard sensors and online learning. Specifically, our work focuses on quadruped locomotion under varying load conditions. This paper presents an adaptive force control framework for quadruped robots, enhanced with online system identification, to handle significant changes in both mass and center of mass (CoM). The proposed approach demonstrates superior velocity and height tracking, even under extreme load conditions, showing promise for applications in logistics, military, and rescue missions.

Keywords: Linear systems, Robust control
Abstract: For a given plant, closed loopshaping (CLS) aims to synthesize a feedback controller under separate constraints on the H-infinity norms of the weighted sensitivity and weighted complementary sensitivity functions. Mixed-sensitivity synthesis (MSS) attempts to perform CLS by constraining the H-infinity norm of the vector formed by stacking these two functions. It is known that MSS is therefore conservative compared to CLS. However, a classical-control-centric graphical characterization of this conservativeness seems to be absent. For single-input single-output plants, this paper provides this characterization, and therefrom presents a nonconservative approach to CLS.

Keywords: Robust control, LMIs, Mechatronics
Abstract: This paper improves a recently appeared sliding mode control methodology for underactuated systems by translating some of its design conditions into numerically testable linear matrix inequalities, which can be solved in polynomial time via commercially available software. In this way, the original methodology which is based on the cascade normal form as well as a set of error variables that define the sliding surface, earns systematicness and flexibility as many other performance specifications can be added. The advantages of the proposal are put at test in a real-time implementation for the inertia pendulum.

Keywords: Robust control, Control applications, Chaotic systems
Abstract: This paper proposes a new non-singular integral terminal sliding mode control (NIT-SMC) for a class of second-order chaotic systems. To resolve the singularity issues in the integral terminal sliding mode control (IT-SMC), the sliding surface is redesigned and a nonsingular control law is proposed. A super-twisting sliding mode observer (STO) is used to observe the unmeasured state and disturbance of the system. The stability of the proposed controller is analysed using the Lyapunov stability concept. The performance of the proposed super-twisting observer-based non-singular integral terminal sliding mode controller (STO-NIT-SMC) is tested with the second-order Duffing chaotic system. Tracking of the desired square and sinusoidal trajectories using the proposed NIT-SMC is successfully achieved, and its superiority is established by comparing the performance with the IT-SMC and the existing NIT-SMC.

Keywords: Robust control, Nonlinear systems, Mechanical systems/robotics
Abstract: Rehabilitation exoskeleton robots are vital for restoring lower limb functionality in individuals with locomotor disorders. Extensive research has focused on optimizing the control of these robotic systems and make them more stable. However, there is still a considerable gap in developing robust controllers. This work introduces a robust control approach tailored for a one-degree-of-freedom (1-DoF) knee rehabilitation exoskeleton robot. Emphasizing position control, our method addresses challenges such as parameter uncertainties, solid and viscous frictions, external disturbances, and motion constraints. To effectively manage these issues, we propose a composite controller that integrates a linear state-feedback controller and a nonlinear control law. By employing a quadratic Lyapunov function, we derive the expression of the nonlinear controller and establish the LMI conditions required for calculating the matrix gain of the composite controller using two different design methods. To confirm the developed LMI conditions and demonstrate the proposed control method's ability to ensure robust position control of the knee exoskeleton robotic system, simulation results are finally provided.

Keywords: Robust adaptive control, Aerospace, Modeling and simulation
Abstract: This paper introduces an improved control strategy for tracking the altitude and attitude of the quadcopter unmanned aerial vehicle (UAV) in the presence of external disturbance, known as adaptive super-twisting terminal sliding mode control (AST-TSMC). This approach ensures finite-time convergence of the sliding variable and error trajectories using a real second-order sliding mode. An adaptive super-twisting approach is incorporated to reduce the reliance on the upper bound of the unknown disturbance and achieve continuous control. A nonlinear sliding surface is designed to ensure the finite time convergence of the sliding variable and the tracking error. The Lyapunov-based stability analysis is presented to confirm the finite-time convergence. Comparative simulation results of AST-TSMC, ST-TSMC, and TSMC are presented to demonstrate the strategy's performance. The effectiveness of the proposed approach is validated through MATLAB simulation results.

Keywords: Reduced order modeling, Linear systems, Computational methods
Abstract: In this work, a novel configuration of frequency-limited controllability and observability Gramians is constructed to develop a new truncated balanced realization for minimizing the error of higher-order models in specified frequency intervals. The suggested technique generates stable simplified models and approximates better than existing techniques. Two numerical examples are used to assess the performance of the presented approach, and it is found through the error plots and comparison that the proposed reduced models approximate the actual models satisfactorily in the specified frequency band.

Keywords: Machine learning, Optimization, Simulation
Abstract: Design optimization of solar photovoltaic (PV) cells is challenging due to the complex interplay of numerous parameters and the computationally intensive nature of high fidelity simulations. This work presents a novel machine learning (ML)-surrogate hybrid optimization framework to address this challenge. Our approach leverages an ML model trained on data generated from a solar PV cell simulator to rapidly predict cell performance based on design parameters. This ML model, a computationally efficient surrogate for the simulator, is then integrated with an optimization algorithm to identify optimal design parameters that maximize the desired PV cell performance metric. The proposed approach is validated through extensive experiments using Quokka 2 as the solar PV cell simulator. Results show that the proposed approach maximizes the chosen cell performance to within 2.1% of that obtained by a rigorous baseline based on Bayesian Optimization, but with a significantly shorter convergence time (∼ 4 times faster).

Keywords: Reduced order modeling, Neural networks, Machine learning
Abstract: Froth flotation is a widely used technique for mineral separation, but it poses significant challenges for modeling due to the complex phenomena that happen during the process. To account for the implementation of optimization and control-related studies in flotation systems, this work deals with development of a digital twin or surrogate to the original model for the froth flotation system. Initially, a first-principles, three-phase model having 61 equations was considered and implemented. The model consists of four inputs, a set of initial conditions for feed, one primary output, and sixteen intermediate outputs. 15,000 time-series data were sampled from the physics-based model to initiate the data-driven modeling. The well-known recurrent neural network variants, namely RNN, LSTM, and GRU, were developed as potential surrogates for the three-phase model by following a multi-input and single-output approach. On performing a comparison study for all the output predictions on test data, it was found that GRU outperforms LSTM and RNN for a majority of the outputs (9 out of 17), especially for highly dynamic ones. Also, GRU consists of 25% fewer model parameters than LSTM networks, which thereby decreases the time for training and reduces the load on the optimizer. Overall, this study provides a deep understanding of froth flotation dynamics and establishes the efficacy of advanced neural network architectures in enhancing predictive modeling in mineral separation processes.

Keywords: Machine learning, Modeling and simulation, Neural networks
Abstract: In a multi-robot system, two or more robots are trained to cooperate in their actions to complete a given task faster and with a higher probability of success than in a single-robot system. To test the performance of these approaches in dynamic situations, we simulated a 3D fire disaster environment set in a real-world urban city building with humans trapped inside it. We model the fire dynamics such that the intensity and spread rate of fire, smoke, and blocked paths increase non-linearly. The trapped humans have a high risk of losing their lives if not searched within a deadline. We state this problem as the Multi-robot Cooperative Search in Dynamic Environments (MCSDE) problem.

The existing methods are not time efficient in solving the MCSDE problem. To address the issue of multi-robot training in MCSDE, we present a modular training approach. In this approach, we train each robot locally based on their local perceptions. Our efforts in this research are to uncover these perceptions' geometric locality properties. When fused with another robot's perception, each perception reveals the true behaviour of the environment's dynamic. Each robot becomes self-aware of global dynamics by training from its local perceptions. The proposed modular approach requires just one-fourth of the computation than the RL or DL-based training approaches and performs far superior to the algorithmic approaches.

Keywords: Modeling and simulation, Aerospace, Algebraic/geometric methods
Abstract: In this paper, we propose a method by which halo orbits in space can be modelled by using Fourier series. We adopted the idea of approximating different 2D and 3D convex and non-convex shapes using Fourier series approximations. The proposed method effectively addresses the challenge of matching not just the shape but also the speed variation of the spacecraft on the halo orbits. We also show that the number of parameters used is reasonably small, thus potentially leading to enormous savings in computation time while providing near perfect halo orbit trajectories.

Keywords: Reduced order modeling, Linear systems
Abstract: This paper presents a novel frequency-weighted model reduction technique for decoupled denominator-type two-dimensional discrete-time systems using truncated balanced realization that can be applied to single- and double-sided weights. The weighted controllability and observability Gramian matrices are block-diagonalized to find the transformation matrix that converts the system to a balanced form. A numerical example of a (3, 3) order first-quadrant Gaussian two-dimensional scalar filter is examined in order to demonstrate the effectiveness of the proposed method.