Real-time Control Performance Research Articles

Augmentation in Performance of Real-Time Balancing and Position Tracking Control for 2-DOF Ball Balancer System Using Intelligent Controllers

Augmentation in Performance of Real-Time Balancing and Position Tracking Control for 2-DOF Ball Balancer System Using Intelligent Controllers

A Real-Time Control Policy to Achieve Maximum Throughput of an Online Order Fulfillment Network

Several major companies operate large online order fulfillment systems to ship goods from fulfillment centers through a distribution network to customer destinations in response to purchase orders. These networks make several types of decisions in real-time to serve customers. First, when a customer places an order, when and where (which fulfillment center) is it fulfilled from? Second, once an order has been packaged, how is it moved through the network to get to the customer? Making optimal decisions can yield significant cost savings or improvements in customer service. Unfortunately, these are large optimization problems, and are furthermore subject to uncertainty in the products and destinations of customer orders and the inventory replenishment of the fulfillment centers. This uncertainty makes the problem difficult to solve to optimality. Although the problem can be modeled as a Markov decision process, solving it exactly using standard computational methods is not possible due to the curse of dimensionality. We propose an alternative approach to this problem. We define a relatively simple real-time control policy and prove that it serves all customer demand if at all possible. This proof is achieved using Lyapunov drift techniques to relate the real-time control performance to the average performance necessary to serve all customers on average. Correspondingly, we characterize the average network performance, which may be used for network topology design while the control policy adapts to real-time stochasticity. We demonstrate the performance and stability properties on a numerical example based on hundreds of Amazon facility locations in the United States. The max-pressure control and greedy policies perform similarly at low demands, but at higher demand the throughput properties of the max-pressure control manifest as improvements in throughput and customer service metrics. Funding: Financial support from the National Science Foundation [Grant 1935514] is gratefully acknowledged. Supplemental Material: The e-companion is available at https://doi.org/10.1287/trsc.2023.0096 .

A Fused Deep Fuzzy Neural Network Controller and Its Application to Pneumatic Flexible Joint

With multilayered deep neural networks (DNNs) distinguished from the traditional single-layer neural networks (SNNs) by the depth, deep learning is a powerful paradigm possessing the capability of feature learning from big data. However, as a completely deterministic model, deep learning does not perform well in dealing with uncertain information and must be pretrained before being applied, which hinders the broad application of deep learning in real-time control. In this article, we introduce the knowledge-based fuzzy logic system (FLS) into deep learning and propose a fused deep fuzzy neural network (FDFNN) controller. The FDFNN controller derives the control signal from both deep fuzzy neural networks and the knowledge-based FLS. At the beginning of control, the FDFNN controller performs like a traditional fuzzy logic controller (FLC), which eliminates the pretraining procedure of deep learning. With another FLS providing supervised signal, the network weights of DFNN's hidden layers can be updated during the control process, which endows the FDFNN controller with online learning capability. A Lyapunov-based convergence analysis is conducted to guarantee the stability of the learning procedure. Both numerical simulations and real-world experiments validate that the proposed method effectively improves the real-time tracking control performance of a pneumatic flexible joint system.

Experimental study of a novel quasi-active negative stiffness damper system for achieving optimal active control performance

Compared to the semi-active control and passive control systems, an active control system is able to achieve the most desirable vibration control performance because of its high adaptability and capacity to directly produce the required optimal control force (ROCF). However, the disadvantages of high energy consumption and low reliability limit its wide application. To address this dilemma, a novel ‘Quasi-active’ control concept and a representative quasi-active negative stiffness damper (QANSD) system integrating a negative stiffness element and an MR damper were proposed by the authors very recently. The QANSD system has been theoretically and numerically verified to be able to fully generate the ROCF and realize active control performance with much less external power consumption than a semi-active control system. To further validate the effectiveness of the innovative concept and the proposed QANSD system, in this study, a comprehensive experimental investigation was conducted based on the developed prototype device. The fundamentals of the QANSD system including realizations and mechanical behaviours are firstly introduced, and the experimental prototype and setup are then provided. Following that, experimental tests on the QANSD system and its MR damper and negative stiffness element components were carried out under different harmonic excitations and command currents. Based on the experimental test results, artificial neural network-based forward and inverse models are designed for the QANSD force and command current identifications. Moreover, the real-time vibration control performances of the QANSD system are investigated and compared with the active and semi-active control systems. The experimental results revealed that compared to the MR damper that was only able to generate force in the opposite direction of structural motion, the developed QANSD system had the capacity to provide both velocity-resisting and velocity-assisting forces as the ROCF does. Its vibration control performance could fully match that of the desired active control system, but with much lower energy consumption.

The effect of ratios upon improving patient-based real-time quality control (PBRTQC) performance.

Large biological variation hinders application of patient-based real-time quality control (PBRTQC). The effect of analyte ratios on the ability of PBRTQC to improve error detection was investigated. Four single analyte-ratio pairs (alanine aminotransferase [ALT] vs. ALT to aspartate aminotransferase ratio [RALT]; creatinine [Cr] vs. Cr to cystatin C ratio [RCr]; lactate dehydrogenase [LDH] vs. LDH to hydroxybutyrate dehydrogenase ratio [RLDH]; total bilirubin [TB] vs. TB to direct bilirubin ratio [RTB]) were chosen for comparison. Various procedures, including four conventional algorithms (moving average [MA], moving median [MM], exponentially weighted moving average [EWMA] and moving standard deviation [MSD]) were assessed. A new algorithm that monitors the number of defect reports per analytical run (NDR) was also evaluated. When a single analyte and calculated ratio used the same PBRTQC parameters, fewer samples were needed to detect systematic errors (SE) by taking ratios (p<0.05). Application of ratios in MA, MM and EWMA significantly enhanced their ability to detect SE. The influence of ratio on random error (RE) detection depended upon the analytes and PBRTQC parameters, as consistent advantage was not demonstrated. The NDR method performed well when appropriate parameters were used, but was only effective for unilateral SE. Rearrangement of sample order led to a significant deterioration of conventional algorithms' performance, while NDR remained almost unaffected. For analytes with large variation and poor PBRTQC performance, using ratios as PBRTQC indexes may significantly improve performance and achieve better anti-interference ability, providing a new class of monitoring indicators for PBRTQC.

An adaptive control strategy for microgrid secondary frequency based on parameter identification

Complex microgrid structures and time-varying conditions, among other factors, cause problems in the mechanical modeling of microgrids, making model-based controller optimization difficult. Therefore, this study proposed a secondary frequency adaptive control strategy based on parameter identification, which uses an online parameter identification method to identify the parameters in the microgrid in real-time. The identified parameters are then used in the secondary frequency adaptive controller to optimize the real-time controller performance. The proposed method realizes adaptive optimization of the controller in the microgrid operation state and is applied to a microgrid with unknown parameters to adjust the controller parameters. Finally, a simulation experiment was conducted to verify the model accuracy and the frequency regulation effect of the proposed adaptive control strategy

Sliding-surface dynamic control of a continuum manipulator with large workspace

This paper presents the trajectory tracking control of a multi-section continuum and compliant manipulator with large workspace. Fluidic muscles are utilized as actuators in this manipulator and inertial sensors used to measure its shape in configuration space. The robot's large workspace signifies the effects of the inertial, gravitational, and applied external forces on the trajectory tracking control of the robot end-effector. In this study, a model-based sliding-surface controller is proposed by deriving a simplified Pseudo-Rigid Body (PRB) dynamic model, which utilizes modified constant curvature assumptions to describe the couplings between the robot sections and actuators. Simplification steps are proposed for a PRB dynamic model by eliminating negligible dynamic effects to achieve a computationally efficient implementation with real-time control performance. The unmodeled dynamics and the error of model simplifications are considered as system uncertainties and an integral sliding surface with a variable sliding gain is proposed to overcome these uncertain dynamics. The sliding gain considers changes in the upper bounds of uncertainties. The closed-loop system stability in trajectory tracking is mathematically proved based on the Lyapunov stability theorem. Experiments throughout the manipulator's large workspace in challenging configurations including sudden release of unknown loads at the robot tip also highlight the potential of our contributed controller.

Evaluating real-time control of stormwater drainage network and green stormwater infrastructure for enhancing flooding resilience under future rainfall projections

Traditional stormwater control measures are designed to handle system loadings induced by fixed-size storm events. However, climate change is predicted to alter the frequency and intensity of flooding events, stimulating the need to explore another more adaptive flooding solution like real-time control (RTC). This study assesses the performance of RTC to mitigate impacts of climate change on urban flooding resilience. A simulated, yet realistic, urban drainage system in Salt Lake City, Utah, USA, shows that RTC improves the flooding resilience by up to 17% under climatic rainfall changes. Compared with green stormwater infrastructure (GSI), RTC exhibits a lower resistibility, lower flooding failure level, and higher recovery rate in system performance curves. Results articulate that keeping RTC's performance consistent under ‘back-to-back’ storms requires a tradeoff between upstream dynamical operation and downstream flooding functionality loss. This research suggests that RTC provides a path towards smart and resilient stormwater management strategy.

An integrated path-tracking and control allocation method for autonomous racing electric vehicles

In recent years, path-tracking controllers for autonomous passenger vehicles and Control Allocation (CA) methods for handling and stability control have both received extensive discussion in the literature. However, the integration of the path-tracking control with CA methods for autonomous racing vehicles has not attracted much attention. In this study, we design an integrated path-tracking and CA method for a prototype autonomous racing electric vehicle with a particular focus on the maximising the turning speed in tight cornering. The proposed control strategy has a hierarchical structure to improve the computational efficiency: the high-level path-tracking Model Predictive Control (MPC) based on a rigid body model is designed to determine the virtual control forces according to the desired path and desired maximum velocity profile, while the low-level CA method uses a Quadratically Constrained Quadratic Programming (QCQP) formulation to distribute the individual control actuator according to the desired virtual control values. The proposed controller is validated in a high-fidelity simulation vehicle model with the computational time of the optimisation controller presented to demonstrate the real-time control performance.

실내 이산화탄소 농도 예측 기반 적응형 환기 제어 알고리즘 개발

Purpose: This study aimed to develop an adaptive ventilation control algorithm for occupant-centric control (OCC). The control algorithm utilizes a real-time indoor carbon dioxide concentration prediction model that reflects occupant information and is continuously updated through daily learning. Method: The prediction model was developed using a long short-term memory (LSTM) learning algorithm based on data obtained from a living lab. The indoor CO2 concentration after 5 minutes was predicted through the data of the past 1 hour, and the prediction accuracy was evaluated with the test data. The adaptive ventilation control algorithm, which incorporates the prediction model, was then applied to the living lab for experiments to evaluate its real-time prediction accuracy, adaptability, and control performance. Result: As a result of the performance evaluation of the predictive model, the coefficient of variation of the root mean squared error (CVRMSE) was 1.78% and the R2 was 0.97. The adaptability evaluation over four days presented an improvement in CVRMSE from 1.78% to 1.13%, which is approximately 36.52% improvement from the initial performance. During the experiment with the adaptive ventilation control algorithm, the accuracy decreased slightly with a CVRMSE of 2.90% and R2 of 0.98, likely due to frequent ventilation control leading to large data variations. Despite short period of the indoor carbon dioxide concentration exceeding 1,000 ppm, the control was effective. According to the results, it is expected that providing comfortable indoor air quality at all times can be achieved by improving the optimal control cycle and supplementing data learning for various control modes in future research.

Closing the Planning–Learning Loop With Application to Autonomous Driving

Real-time planning under uncertainty is critical for robots operating in complex dynamic environments. Consider, for example, an autonomous robot vehicle driving in dense, unregulated urban traffic of cars, motorcycles, buses, etc. The robot vehicle has to plan in both short and long terms, in order to interact with many traffic participants of uncertain intentions and drive effectively. Planning explicitly over a long time horizon, however, incurs prohibitive computational cost and is impractical under real-time constraints. To achieve real-time performance for large-scale planning, this work introduces a new algorithm <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Learning from Tree Search for Driving</i> (LeTS-Drive), which integrates planning and learning in a closed loop, and applies it to autonomous driving in crowded urban traffic in simulation. Specifically, LeTS-Drive learns a policy and its value function from data provided by an online planner, which searches a sparsely sampled belief tree; the online planner in turn uses the learned policy and value functions as heuristics to scale up its run-time performance for real-time robot control. These two steps are repeated to form a closed loop so that the planner and the learner inform each other and improve in synchrony. The algorithm learns on its own in a self-supervised manner, without human effort on explicit data labeling. Experimental results demonstrate that LeTS-Drive outperforms either planning or learning alone, as well as open-loop integration of planning and learning.

A Novel STFSA-CNN-GRU Hybrid Model for Short-Term Traffic Speed Prediction

Short-term traffic speed prediction is fundamental to intelligent transportation systems (ITS), and the accuracy of the model largely determines the performance of real-time traffic control and management. In this study, a short-term traffic speed prediction method based on the spatial-temporal analysis of traffic flow and a combined deep-learning model, and a hybrid spatial-temporal feature selection algorithm (STFSA) of a convolutional neural network–gated recurrent unit (CNN-GRU)) is initially developed. Specifically, the STFSA is firstly employed to reconstruct the spatial-temporal matrix of traffic speed based on temporal continuity and spatial characteristics, and then this matrix is considered as the input feature of the prediction model. After this, the nonlinear fitting ability of the CNN is adopted to extract deep features from the convolutional and pooling layers for model training. Finally, by combining the timing and long-range dependence of the captured data with the forward GRU and the reverse GRU, the accuracy of the prediction result is further improved. The validity of the proposed model can be verified by comparing the prediction results with the actual traffic data. Accordingly, in the case study, the performance is compared with various benchmark methods under the same prediction scenario, verifying the superiority of the proposed model.

Toward Multi-Service Edge-Intelligence Paradigm: Temporal-Adaptive Prediction for Time-Critical Control over Wireless

Time-critical control applications typically pose stringent connectivity requirements for communication networks. The imperfections associated with the wireless medium such as packet losses, synchronization errors, and varying delays have a detrimental effect on performance of real-time control, often with safety implications. This paper introduces multi-service edge-intelligence as a new paradigm for realizing time-critical control over wireless. It presents the concept of multi-service edge-intelligence which revolves around tight integration of wireless access, edge-computing and machine learning techniques, in order to provide stability guarantees under wireless imperfections. The paper articulates some of the key system design aspects of multi-service edge-intelligence. It also presents a temporal-adaptive prediction technique to cope with dynamically changing wireless environments. It provides performance results in a robotic teleoperation scenario. Finally, it discusses some open research and design challenges for multi-service edge-intelligence.

Principle and control strategy of a novel wave-to-wire system embedded ocean energy storage optimization

Fluctuation and unpredictability of wave power output affect the safe operation of the power grid, which greatly restricts the development of wave power generation. This paper firstly introduces the principle and control strategy of wave to wire (W2W) model, whose input is from wave energy conversion (WEC) system and output to the electrical power injected into grid for stable power. Then, an adapted particle swarm optimization (APSO) algorithm-based maximum power point tracking technology (APSO-MPPT) considering adaptive inertia weighting variables is presented for quick dynamic reaction. Comparisons are made with the traditional constant voltage method, INC method, P&Q method, and PSO algorithm, the APSO-MPPT method can reduce the tracking time by 9.3% less than the quickest traditional. Secondly, a K–SOC curve algorithm-based hybrid energy storage system (KSC-HESS) optimization is conducted for smoothing the wave power output, with real time control performance. Simulations under output power fluctuation, load pulsation and load unbalancing situations are occupied for demonstrating high performance for the characteristics of the proposed KSC-HESS, in both reducing high-frequency burden for super-capacitors and protecting batteries from over voltage. Furthermore, results of prototype experiment verified the effectiveness with better efficiency by adapting the proposed APSO-MPPT and KSC-HESS for the proposed W2W system. Analysis shows the proposed W2W approach has comparative advantages of control complexity and high reliability for renewable power output.

Towards coordinated and robust real-time control: a decentralized approach for combined sewer overflow and urban flooding reduction based on multi-agent reinforcement learning

The real-time control (RTC) of urban drainage systems can make full use of the capabilities of existing infrastructures to mitigate combined sewer overflow (CSO) and urban flooding. Despite the benefits of RTC, it may encounter potential risks and failures, which need further consideration to enhance its robustness. Besides failures of hardware components such as sensors and actuators, the RTC performance is also sensitive to communication failures between the devices that are spatially distributed in a catchment-scale system. This paper proposes a decentralized control strategy based on multi-agent reinforcement learning to enhance communication robustness and coordinate the decentralized control agents through centralized training. To investigate different control structures, a centralized and a fully decentralized strategy are also developed based on reinforcement learning (RL) for comparison. A benchmark drainage model and a real-world drainage model are formulated as two cases, and the control agents are trained to control the orifices or pumps for CSO or flooding mitigation in each case. The three RL strategies reduce the CSO volume by 5.62-9.30% compared with a static baseline in historical rainfalls of the benchmark case and reduce the CSO and flooding volume by 14.39-21.36% compared with currently-used rule-based control in synthetic rainfalls of the real-world case. Benefitting from centralized training, the decentralized agents can achieve similar performance to the centralized agent. The decentralized control also enhances the communication robustness with smaller performance loss than the centralized control when observation communication fails, and provides a robust backup at the local level to limit the uncertainties when action commands from the centralized agent are lost. The results and findings indicate that multi-agent RL contributes to a coordinated and robust solution for RTC of urban drainage systems.

Dynamic modeling of vanadium redox flow batteries: Practical approaches, their applications and limitations

Vanadium redox flow batteries (VRFBs) have been in the focus of attention of the energy storage community over the past years. Adequate, reliable and user-friendly mathematical models are required for the development and optimal application of this type of battery. A large amount of literature has been devoted to dynamic models of VRFBs, but insufficient attention has been paid to the comparison and critical analysis of their applicability. The article provides a comprehensive overview of available dynamic models, comparing their applicability for real-time simulation of industrial-scale vanadium redox flow batteries. In addition, a methodology for the models comparison is proposed, which takes into account different modes of operation and allows to determine the applicability range of a particular model. The results of the study show that lumped-parameter models with crossover performs well for simulating battery dynamics in the wide range of operating conditions: state of charge (SOC) – (0.1–0.8), load current – (20–290 mA/cm2), flow velocity - (0.4–2.7 cm/s). Also, such models can be tuned even if some physical parameters of the battery components (e.g. electrodes and membrane) are unknown. The results obtained can be used to provide more accurate simulation of vanadium redox flow batteries in real-time monitoring and control tasks, when accuracy and performance are important.

퍼블릭 클라우드 기반 셀프 밸런싱 로봇 네트워크 제어 시스템

In a networked control system, a controlled model and a controller are remotely operated via network communication. In this paper, we propose a public cloud-based real-time networked control system and demonstrate the feasibility of its real-time control performance by building a self-balancing robot and conducting experiments with it. We first build the self-balancing robot as the controlled model and implement the controller into a docker container for the networked control system. In addition, we utilize the google cloud platform as the computing and networking resources to establish Kubernetes platform, which operates the control system. Through the experiments, we show the self-balancing robot is controlled and maintains its balance in the cloud-based networked control system.

Self-balancing collaboration between linear and nonlinear adaptive sub-filters for hybrid active noise control

Conventional adaptive filters for active noise control (ANC) are troubled by the trade-off between convergence speed and steady state error, especially when nonlinearity plays an important role in system model. In this paper, a hybrid ANC algorithm called FBFLANN is constructed by the collaboration of FIR and bilinear functional link artificial neural network. The characteristics of linear and nonlinear sub-filters are analyzed and utilized to facilitate the adaptability in noise reduction. A convex factor is employed to automatically balance the contribution of two sub-filters, achieving faster convergence and higher accuracy in the presence of nonlinear distortions. The mathematical working and updating rules of the controller are derived to regulate the ANC process theoretically. The effectiveness of the proposed FBFLANN algorithm is demonstrated by numerous simulation studies considering diverse nonlinear acoustic path models as well as acoustic feedback interference. Hardware ANC platform is built up to verify the real-time noise control performance in physical noisy environments. The simulation and experiment results confirm the applicability and prospects for further development of the presented architecture in noise control scenarios.

Policy Search for Model Predictive Control With Application to Agile Drone Flight

Policy search and model predictive control (MPC) are two different paradigms for robot control: policy search has the strength of automatically learning complex policies using experienced data, and MPC can offer optimal control performance using models and trajectory optimization. An open research question is how to leverage and combine the advantages of both approaches. In this article, we provide an answer by using policy search for automatically choosing high-level decision variables for MPC, which leads to a novel <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">policy-search-for-model-predictive-control framework</i> . Specifically, we formulate the MPC as a parameterized controller, where the hard-to-optimize decision variables are represented as high-level policies. Such a formulation allows optimizing policies in a self-supervised fashion. We validate this framework by focusing on a challenging problem in agile drone flight: flying a quadrotor through fast-moving gates. Experiments show that our controller achieves robust and real-time control performance in both simulation and the real world. The proposed framework offers a new perspective for merging learning and control.

Layer-by-layer model-based adaptive control for wire arc additive manufacturing of thin-wall structures

Improving the geometric accuracy of the deposited component is essential for the wider adoption of wire arc additive manufacturing (WAAM) in industries. This paper introduces an online layer-by-layer controller that operates robustly under various welding conditions to improve the deposition accuracy of the WAAM process. Two control strategies are proposed and evaluated in this work: A PID algorithm and a multi-input multi-output model-predictive control (MPC) algorithm. After each layer of deposition, the deposited geometry is measured using a laser scanner. These measurements are compared against the CAD model, and geometric errors are then compensated by the controller, which generates a new set of welding parameters for the next layer. The MPC algorithm, combined with a linear autoregressive (ARX) modelling process, updates welding parameters between successive layers by minimizing a cost function based on sequences of input variables and predicted responses. Weighting coefficients of the ARX model are trained iteratively throughout the manufacturing process. The performance of the designed control architecture is investigated through both simulation and experiments. Results show that the real-time control performance is improved by increasing the complexity of implemented control algorithm: controlled geometric fluctuations in the test component were reduced by 200% whilst maintaining fluctuations within a 3 mm limit under various welding conditions. In addition, the adaptiveness of designed control strategy is verified by accurately controlling the fabrication of a part with complex geometry.