In this paper we address the simultaneous collision detection and force estimation problem for quadrupedal locomotion using joint encoder information and the robot dynamics only. We design an interacting multiple-model Kalman filter (IMM-KF) that estimates the external force exerted on the robot and multiple possible contact modes. The method is invariant to any gait pattern design. Our approach leverages pseudo-measurement information of the external forces based on the robot dynamics and encoder information. Based on the estimated contact mode and external force, we design a reflex motion and an admittance controller for the swing leg to avoid collisions by adjusting the leg's reference motion. Additionally, we implement a force-adaptive model predictive controller to enhance balancing. Simulation ablatation studies and experiments show the efficacy of the approach.

Limited data has become a major bottleneck in scaling up offline imitation learning (IL). In this paper, we propose enhancing IL performance under limited expert data by introducing a pre-training stage that learns dynamics representations, derived from factorizations of the transition dynamics. We first theoretically justify that the optimal decision variable of offline IL lies in the representation space, significantly reducing the parameters to learn in the downstream IL. Moreover, the dynamics representations can be learned from arbitrary data collected with the same dynamics, allowing the reuse of massive non-expert data and mitigating the limited data issues. We present a tractable loss function inspired by noise contrastive estimation to learn the dynamics representations at the pre-training stage. Experiments on MuJoCo demonstrate that our proposed algorithm can mimic expert policies with as few as a single trajectory. Experiments on real quadrupeds show that we can leverage pre-trained dynamics representations from simulator data to learn to walk from a few real-world demonstrations.

This paper proposes a novel physics-informed ma-chine learning framework for motion planning and control of autonomous vehicles. By integrating longitudinal and lat-eral control, a nonlinear control problem is formulated using Model Predictive Control (MPC). To address computational challenges, a self-supervised framework, Recurrent Predictive Control (RPC), is introduced, leveraging differentiable neural networks and recurrent neural networks to train a neural network controller. Additionally, a heuristic feedback control layer is designed to reduce steady-state errors in the closed-loop tracking. Through numerical simulations and co-simulations using Simulink and CarSim, five neural network controllers are compared with an MPC controller in a lane-changing sce-nario. The proposed RPC framework improves computational efficiency by 95% compared to MPC, enhances generalization performance compared to Approximate MPC, and reduces performance loss by 17% compared to Differentiable Predictive Control. The heuristic feedback control layer further reduces steady-state errors and improves convergence speed during training.

Hybrid cloud-edge systems combine the advantages of cloud computing and mobile edge computing (MEC) to achieve flexible integration and fluidity of data between the cloud and the edge. To address dynamic and stochastic loads caused by mobile users (MUs) and time-varying tasks, MEC network operators need to continuously migrate installed services among edge servers, significantly increasing network maintenance costs. Existing studies often overlook the service migration cost resulting from MU mobility. Therefore, we present a joint optimization scheme focusing on minimizing the operational cost of hybrid cloud-edge systems while considering the dynamic service migration cost induced by MUs. With the rapid development of 5G/6G technologies, many MUs require connectivity to edge nodes (ENs) or cloud data centers (CDCs) for processing. Minimizing the operational cost of hybrid cloud-edge systems while considering many heterogeneous decision variables is a challenge. To solve this complex high-dimensional mixed-integer nonlinear problem, we develop a novel deep learning-based evolutionary algorithm called autoencoder-based multiswarm gray wolf optimizer based on genetic learning (AMGG). Experimental results with real data demonstrate that AMGG achieves lower system cost by 49.69% while strictly meeting task latency requirements of MUs compared with state-of-the-art algorithms.

Along with the fourth industrial revolution, industrial automation systems are evolving into a multi-tier end-edge computing architecture. Edge controllers, which are equipped with a larger computing capacity compared to local controllers, can communicate with local plants over mainstream wireless networks such as WirelessHART, Wi-Fi, and cellular networks. Well-known challenges induced by networks, such as uncertain time delays and packet drops, have been intensively investigated from various perspectives: control synthesis, network design, or control and network co-design. The status quo is that the industry remains hesitant to close the loop between the edge controller and the actuation side due to safety concerns. This work offers an alternative perspective to address the safety concern, by exploiting the design freedom of an end-edge computing architecture. Specifically, we present a smart actuation framework, which deploys (1) an edge controller, which communicates with physical plant via wireless network, accounting for optimality, adaptation, and constraints by conducting computationally expensive operations; (2) a smart actuator, which is co-located with the physical plant on the end tier and executes a local control policy, accounting for system safety in the view of network imperfections, (3) the end-edge control co-design strategies and cooperation logic for both performance and stability. For certain classes of plants, semi-globally asymptotic stability of the resulting end-edge control systems is established when the edge controller is the model predictive control (MPC), or policy iteration-based learning control. We also provide an adaptation strategy for the end-edge control systems facing model parameter mismatches when the edge controller employs reinforcement learning. Extensive simulations demonstrate the advantages of the proposed end-edge co-design and cooperation procedures. Note to Practitioners—Edge computing is gaining momentum in areas that require low latency and high efficiency, i.e., mobile computing, video analytics, and autonomous driving. Industrial automation systems are also evolving into a multi-tier end-edge computing architecture. It pays obvious dividends to leverage the cooperation between end and edge, benefiting from fast and reliable communication on the end side, and powerful computation capacity on the edge side. The current end-edge cooperation focuses on how to partition tasks and offload computation resources in order to minimize delay and energy consumption, as well as how to balance the tradeoff between them. However, the impacts of end-edge cooperation on the safety, optimality, and cost of industrial automation have not been systematically studied. This paper aims to tailor end-edge cooperation in a smart actuation framework, for industrial automation to reconcile the above aspects by leveraging co-design of end and edge controllers and their switching logic. Extensive pure and semi-physical simulations demonstrate the advantages in performance and system stability of the proposed end-edge co-design and cooperation procedures.

Isolated converters are the ideal candidate for high-gain DC/DC converter applications. Although the design of converters considers safety margins and derating requirements to ensure acceptable performance until the end of product life, it does not completely rule out any early component failure. It is inevitable that the post-deployment stresses result in component aging and subsequent failure before the end of its useful life. To minimize downtime, early failure signatures need be detected for predictive maintenance, which requires continuous/periodic monitoring of critical components. This paper presents an Extended Kalman Filter (EKF)-based predictive maintenance algorithm of a 2kW, 270V/28V DC/DC converter. The detailed mathematical framework and observability analysis highlighting the feasibility of state/parameter estimation are presented. Based on observability analysis, the converter operating condition is devised such that accurate parameter estimation is achievable at a lower sensor sampling rate (equal to the switching time period). The significance of module-exclusive measurements on the observability and estimation of parameters is also discussed for a two-module system. The analyses presented are validated with detailed simulation results for single (2kW) and double ($2\times 2\text{kW}$) modules to estimate parameters of the filter inductor and the output capacitor.

Aviation industry is moving towards more electric aircraft, where both non-propulsive and propulsive loads are electrified. While bringing in various benefits, electric propulsion system (EPS) introduces extra complexity and weight to aircraft, as well as raises the reliability concern. New design and analysis tools are required to size the EPS while meeting stringent reliability requirements. This paper investigates how to consolidate readability into the EPS sizing process and makes two-fold contributions. First, a probabilistic algorithm is proposed to assess the reliability of the EPS admitting a directed acyclic graph topology. The algorithm reduces the directed acyclic graph to a layered tree, which simplifies the calculation of joint probability of nodes in each layer. Second, we formulate the reliability-based EPS sizing as an integer nonlinear programming problem, where the reliability requirements are posed as constraints. Preliminary simulation validates the proposed method.

This paper presents an integrated motion planning system for autonomous vehicle (AV) parking in the presence of other moving vehicles. The proposed system includes 1) a hybrid environment predictor that predicts the motions of the surrounding vehicles and 2) a strategic motion planner that reacts to the predictions. The hybrid environment predictor performs short-term predictions via an extended Kalman filter and an adaptive observer. It also combines short-term predictions with a driver behavior cost-map to make long-term predictions. The strategic motion planner comprises 1) a model predictive control-based safety controller for trajectory tracking; 2) a search-based retreating planner for finding an evasion path in an emergency; 3) an optimization-based repairing planner for planning a new path when the original path is invalidated. Simulation validation demonstrates the effectiveness of the proposed method in terms of initial planning, motion prediction, safe tracking, retreating in an emergency, and trajectory repairing.

This paper presents a motion planning strategy that utilizes the improved A -search guided tree to enable autonomous parking of a general 3-trailer with a car-like tractor. Different from the state-of-the-art state-lattice-based methods, where numerous motion primitives are necessary to ensure successful planning, our work allows quick off-lattice exploration to find a solution. Our treatment brings at least three advantages: fewer and lower design complexity of motion primitives, improved success rate, and increased path quality. Unlike on-lattice exploration, where the cost-to-go is obtained by querying a heuristic look-up table, off-lattice exploration entails the heuristic function being well-defined at off-lattice nodes. We train a neural network through reinforcement learning to model the maneuver costs of the trailer and use it as the heuristic value to better approximate the cost-to-go. Simulations demonstrate the effectiveness of the proposed method in terms of planning speed and path length.

Data-driven methods for learning optimal control policies such as adaptive dynamic programming have garnered widespread attention. A strong contrast to full-fledged theoretical research is the scarcity of demonstrated successes in industrial applications. This paper extends an established data-driven solution for a class of adaptive optimal linear output regulation problem to achieve self-tuning torque control of servomotor drives, and thus enables online adaptation to unknown motor resistance, inductance, and permanent magnet flux. We make contributions by tackling three practical issues/challenges: 1) tailor the baseline algorithm to reduce computation burden; 2) demonstrate the necessity of perturbing reference in order to learn feedforward gain matrix; 3) generalize the algorithm to the case where F matrix in the output equation is unknown. Simulation demonstrates that the deployment of adaptive dynamic programming lands at optimal torque tracking policies.

This paper presents a hierarchical motion planning approach that can provide real-time parking plans for autonomous vehicles with limited memory. Through combining a high-level route planner that searches for collision-free routes given traffic and obstacle information and a low level motion planner that considers vehicle dynamics, our approach generates smooth trajectories with reasonable parking behaviors rapidly with very low memory consumption. This hierarchical approach allows for online path repairing and replanning when newly detected obstacles that were not indicated on the offline map obstruct the original planned trajectory. It employs a fast clearance checking procedure to obtain a practical indicator of repairability as well as heuristic guidance for rapid trajectory repairing, and utilizes the high-level route planner to conduct real-time replanning when trajectory repairing is deemed to be difficult. Performance analysis on parking tasks in simulation environments demonstrates the advantages of the proposed approach in terms of both trajectory quality and planning time.

The rapidly growing use of lithium-ion batteries across various industries highlights the pressing issue of optimal charging control, as charging plays a crucial role in the health, safety, and life of batteries. The literature increasingly adopts model predictive control (MPC) to address this issue, taking advantage of its capability of performing optimization under constraints. However, the computationally complex online constrained optimization intrinsic to MPC often hinders real-time implementation. This article is thus proposed to develop a framework for real-time charging control based on explicit MPC (eMPC), exploiting its advantage in characterizing an explicit solution to an MPC problem, to enable real-time charging control. This article begins with the formulation of MPC charging based on a nonlinear equivalent circuit model. Then, multisegment linearization is conducted to the original model, and applying the eMPC design to the obtained linear models leads to a charging control algorithm. The proposed algorithm shifts the constrained optimization to offline by precomputing explicit solutions to the charging problem and expressing the charging law as piecewise affine functions. This drastically reduces not only the online computational costs in the control run but also the difficulty of coding. Extensive numerical simulation and experimental results verify the effectiveness of the proposed eMPC charging control framework and algorithm. The research results can potentially meet the needs for real-time battery management running on embedded hardware.

We develop a method for obtaining safe initial policies for reinforcement learning via approximate dynamic programming (ADP) techniques for uncertain systems evolving with discrete-time dynamics. We employ the kernelized Lipschitz estimation to learn multiplier matrices that are used in semidefinite programming frameworks for computing admissible initial control policies with provably high probability. Such admissible controllers enable safe initialization and constraint enforcement while providing exponential stability of the equilibrium of the closed-loop system.

Industrial automation is embracing wireless sensor-actuator networks (WSANs). Despite the success of WSANs for monitoring applications, feedback control poses significant challenges due to data loss and stringent energy constraints in WSANs. Holistic control adopts a cyber-physical system approach to overcome the challenges by orchestrating network reconfiguration and process control at run time. Fundamentally, it leverages self-awareness across control and wireless boundaries to enhance the resiliency of wireless control systems. In this article, we explore efficient holistic control designs to maintain control performance while reducing the communication cost. The contributions of this work are five-fold: (1) We introduce a holistic control architecture that integrates Low-power Wireless Bus (LWB) and two control strategies, rate adaptation and self-triggered control ; (2) We present heuristics-based and optimal rate selection algorithms for rate adaptation; (3) We design novel network adaptation mechanisms to support rate adaptation and self-triggered control in a multi-hop WSAN; (4) We build WCPS-RT, a real-time network-in-the-loop simulator that integrates MATLAB/Simulink and a physical WSAN testbed to evaluate wireless control systems; (5) We empirically explore the tradeoff between communication cost and control performance in holistic control approaches. Our studies show that rate adaptation and self-triggered control offer advantages in control performance and energy efficiency, respectively, in normal operating conditions. The advantage in energy efficiency of self-triggered control, however, may diminish under harsh physical and wireless conditions due to the cost of recovering from data loss and physical disturbances.

With node selection being directed by a heuristic cost [1]–[3], A-search guided tree (AGT) is constructed on-the-fly and enables fast kinodynamic planning. This work presents two variants of AGT to improve computation efficiency. An improved AGT (i-AGT) biases node expansion through prioritizing control actions, an analogy of prioritizing nodes. Focusing on node selection, a bi-directional AGT (BAGT) introduces a second tree originated from the goal in order to offer a better heuristic cost of the first tree. Effectiveness of BAGT pivots on the fact that the second tree encodes obstacles information near the goal. Case study demonstrates that i-AGT consistently reduces the complexity of the tree and improves computation efficiency; and BAGT works largely but not always, particularly with no benefit observed for simple cases.

Online temperature monitoring of electric motors is essential to the safety of the system under dynamic operation. The number of temperature sensors and their locations are often limited due to physical constraints and cost of the hardware, and the temperatures for most parts of a motor cannot be directly measured. To estimate the instantaneous temperature distribution of a motor, we design an observer for the real-time temperature monitoring using a thermal circuit model and limited measurements. Challenges imposed to the observer design include unknown heat sources and measurement noises. The observer needs to be able to simultaneously estimate all the hidden states and unknown inputs, while dealing with measurement noises. In this paper, different observers, including Kalman filter, Luenberger observer, adaptive observer, and modified proportional-derivative (PD) observer are designed to address the problem. We first give some background information of the problem, and introduce the thermal circuit model; then describe the observer designs with a focus on PD observer, which is more recently developed. The proposed observers are then implemented with simulations, and their performances are evaluated and compared.

Data-driven methods for learning near-optimal control policies through approximate dynamic programming (ADP) have garnered widespread attention. In this paper, we investigate how data-driven control methods can be leveraged to imbue near-optimal performance in a core component in modern factory systems: the electric motor drive. We apply policy iteration-based ADP to an induction motor model in order to construct a state feedback control policy for a given cost functional. Approximate error convergence properties of policy iteration methods imply that the learned control policy is near-optimal. We demonstrate that carefully selecting a cost functional and initial control policy yields a near-optimal control policy that outperforms both a baseline nonlinear control policy based on backstepping, as well as the initial control policy.

This work investigates energy-optimal motion planning for a class of multiaxis motion systems where the system dynamics are linear time-invariant and decoupled in each axis. Solving the problem in a reliable and efficient manner remains challenging owing to the presence of various constraints on control and states, nonconvexity in its cost function, and obstacles. This paper shows how the cost function can be convexified by considering the system dynamics, while decomposing decision variables to obtain a convex representation of collision avoidance constraints. With the convexified cost function and constraints, the original problem is decomposed into two quadratic programming (QP) problems. An alternating quadratic programming (AQP) algorithm is proposed to solve both the QP problems alternatingly and iteratively till convergence. Requiring an initial feasible trajectory as a guess, AQP necessarily converges to an energy-efficient solution that is homotopic to the initial guess. Under certain circumstances, AQP is guaranteed to produce a local optimum. Simulation demonstrates that AQP is computationally efficient and reliable while claiming comparable energy saving as the mixed-integer QP approach.

This paper presents a cooperative/shared framework of the driver and his/her semi-autonomous vehicle in order to achieve desired steering performance. In particular, a copilot controller and the driver together operate and control the vehicle. Exploiting the classical small-gain theory, our proposed shared steering controller is developed independent of the unmeasurable internal states of the human driver, and only relies on his/her steering torque. Furthermore, by adopting data-driven adaptive dynamic programming and an iterative learning scheme, the shared steering controller is studied from the measurable data of the driver and the vehicle. Meanwhile, the accurate knowledge of the driver and the vehicle dynamics is unnecessary, which settles the problem of their potential parametric variations/uncertainties in practice. The effectiveness of the proposed method is validated by rigorous analysis and demonstrated by numerical simulations.

This paper investigates the consensus problem of general linear discrete-time multiagent systems by using distributed model predictive control (DMPC) with self-triggered mechanism. First, a novel DMPC-based consensus algorithm is proposed, where each agent only needs to obtain its neighbors' predicted state sequences once at each time step. We prove that the resultant DMPC optimization problem is feasible, and the proposed algorithm guarantees the dynamic consensus of agents. Then, to further reduce the communication cost and the energy consumption of control updates, a self-triggered DMPC-based consensus algorithm is proposed with the control input and the triggering interval jointly optimized. Numerical examples including the benchmark problem with platooning vehicles are provided to verify the effectiveness and advantages of the proposed algorithms.

We consider in this article battery state of power (SoP) estimation, in particular, we propose two algorithms for predicting voltage corresponding to a future current profile that is known to be demanded by the battery load. The proposed algorithms belong to the class of data-driven methods and are based on the Gaussian Process Regression (GPR) framework. In comparison to conventional model-based approaches, data-driven approaches circumvent the issue of observability of SoP from measurements, especially pronounced in batteries with flat open circuit voltage (OCV) characteristic. In addition, the GPR framework admits accurate modeling of a fairly complicated battery dynamics using training data. Finally, the considered setup enables a relatively easy access to training data whenever the necessity for retraining arises, such as due to battery aging. The proposed algorithms aim to handle diverse battery operating conditions involving smooth and abruptly changing voltage/current measurements with both relatively small and large training datasets. The algorithms are tested on two such datasets, and the measured prediction performance and computation time verify their viability for real-time industrial use. We conclude with a number of possible directions for future research.

This paper presents an application of Gaussian process regression (GPR) to estimate a state of charge (SoC) of Lithium-ion (Li-ion) batteries with different kernel functions. One of the practical advantages of using GPR is that uncertainties in the estimates can be quantified, which enables reliability assessment of the SoC estimate. The inputs of GPR are voltage, current and temperature measurements of the battery and the output is an estimate of SoC. First, training is performed in which optimal hyperparameters of a kernel function are determined to model data properties. Then, the battery SoC is estimated online based on the trained model. The kernel function is the key element in the GPR model since it encodes the prior assumptions about the properties of the function being modeled. Therefore, the impact of kernel function selection on the estimation performance is analyzed using both simulated data and experimental data collected from a LiMn2O4/hard-carbon battery with a nominal capacity of 4.93Ah operating under constant charge and discharge currents.

This brief studies the optimal codesign of nonlinear control systems: simultaneous design of physical plants and related optimal control policies. Nonlinearity of the optimal codesign problem could come from either a nonquadratic cost function or the plant. After formulating the optimal codesign into a nonconvex optimization problem, an iterative scheme is proposed in this brief by adding an additional step of system-equivalence-based policy improvement to the conventional policy iteration. We have proved rigorously that the closed-loop system performance can be improved after each step of the proposed policy iteration scheme, and the convergence to a suboptimal solution is guaranteed. It is also shown that under certain conditions, this additional policy improvement step can be conducted by solving a quadratic programming problem. The linear version of the proposed methodology is addressed in the context of linear quadratic regulator. Finally, the effectiveness of the proposed methodology is illustrated through the optimal codesign of a load-positioning system.

This paper presents a generalized space vector modulation (SVM) method for any modular multilevel converter (MMC). The proposed SVM method produces the maximum level number (i.e., 2n+1, where n is the number of submodules in the upper or lower arm of each phase) of the output phase voltages and a higher equivalent switching frequency than other modulation methods, which consequently leads to reduced harmonics in the output voltages and currents. Compared with earlier modulation methods for MMCs, the proposed SVM method provides two more degrees of freedom, i.e., the redundant switching sequences and the adjustable duty cycles, thus offering significant flexibility for optimizing the circulating current suppression and capacitor voltage balancing. This SVM method is a useful tool for further studies of MMCs, as it can be conveniently extended for any control objectives. The demonstrated results validate the analysis.