Distributed state estimation is essential for modern industrial systems, especially in sensor networks and multi-agent frameworks constrained by limited communication and computational resources. This article presents a novel$\ell$-step Distributed Moving Horizon Estimation (DMHE$\ell$) algorithm that explicitly addresses sporadic measurements. The algorithm employs an information diffusion mechanism, enabling each sensor to leverage measurements from its$\ell$-step neighbors. A comprehensive stability analysis proves that the estimation error remains uniformly bounded, thereby establishing convergence guarantees, regardless of the sporadic nature of the measurements. The effectiveness of DMHE$\ell$is demonstrated through simulations in a realistic industrial scenario involving an autonomous mobile robot. The results confirm that the method achieves high estimation accuracy while requiring significantly less communication and data exchange than existing distributed estimators for comparable accuracy levels.

On-site computational capabilities limit the optimization of electricity production in commercial plants. This work addresses the implementation of a virtual operator for electric power maximization in parabolic-trough collector plants by manipulating the flow rate circulating through the pipes with a neural network-based procedure. A model predictive control (MPC) strategy is proposed using nonlinear models to predict the system’s response. The effect of including diverse parts of the plant in the prediction model and using different objective functions (the model of the pump and the pipes and the difference between optimizing thermal and electric power) is analyzed. First, two control layers are implemented: one for obtaining the temperature setpoints and one for computing the flow rate. Since the nonlinear MPC cannot be computed in real-time for medium and large plants, an artificial neural network is trained to learn the optimal solution and lower the computational burden, reducing a 2-layer MPC strategy into only one neural controller that internally decides the operating point. The simulation results obtain a mean time reduction of 99.99995%, while the electricity production for the studied cases is only reduced by around 1%, making the controller implementable in real-time in actual plants.

This article proposes a real-time implementation of distributed model predictive controllers to maximize the thermal energy generated by parabolic trough collector fields. For this control strategy, we consider that each loop of the solar collector field is individually managed by a controller, which can form coalition with other controllers to attain its local goals while contributing to the overall objective. The formation of coalitions is based on a market-based mechanism in which the heat transfer fluid is traded. To relieve the computational burden online, we propose a learning-based approach that approximates optimization problems so that the controller can be applied in real time. Finally, simulations in a 100-loop solar collector field are used to assess the coalitional strategy based on neural networks in comparison with the coalitional model predictive control. The results show that the coalitional strategy based on neural networks provides a reduction in computing time of up to 99.74% and a minimal reduction in performance compared to the coalitional model predictive controller used as the baseline.

This paper discusses the application of coalitional model predictive control (MPC) to freeways traffic networks, where the goal is reducing the time spent by the drivers through a dynamic setting of variable speed limits (VSL) and ramp metering. The prediction model METANET is used to represent the traffic flows evolution. The system behavior and objective function lead to a non-convex and non-linear optimization problem, which can only be solved in a centralized fashion for small networks. The underlying motivation of this paper is the continued advance of clustering methods in the control of large-scale and spatially distributed systems. The global freeway system is partitioned into a set of coupled sub-stretches, which in turn are assigned to the different agents involved in the control problem. These local controllers can dynamically assemble into coalitions to take coordinated measures. In this work, a top-down approach is considered: the bottom layer consists of the set of controllers that compute the VSL and ramp-metering across time; and the supervisory layer changes periodically the information exchange structure to promote coalitions of those controllers that bring greater performance to the global system. In this way, a balance is sought between optimality and efficiency. Finally, the coalitional approach is simulated on a stretch of traffic freeway where cooperation with adjacent sub-stretches is allowed.

This paper proposes a macroscopic model and two control algorithms for the dynamic operation of reversible lanes on freeways. The proposed model is an extension of the second-order traffic flow model METANET. The reversible lanes are modeled like variable lane drops (taking into account that the cars in the closed/opened lanes need a certain time to leave/enter the corresponding segments). Based on this model, two kinds of dynamic controllers have been developed. The first one is an easy-to-implement logic-based controller that takes into account the congestion lengths generated by the reversible lane bottleneck and uses this information for the dynamic operation of the lanes. The second one is a discrete model predictive control that minimizes the total time spent of the modeled network within some constraints for the maximum values of the generated bottleneck queues. The discrete optimization is carried out via evaluation of the cost function for all the leaves in a reduced search tree. The proposed model and control algorithms are simulated and tested using loop detector data collected over a section of the SE-30 freeway in Seville, Spain. The modeled network includes the Centenario Bridge, which is a bottleneck with a reversible lane that creates recurrent congestion during the morning rush-hour period. The results show that the proposed model is able to reproduce traffic congestion due to the reversible lanes and that all the proposed controllers (which can be computed in a short time) substantially reduce this congestion.

This work presents the design, implementation and experimental evaluation of the control system for a Fuel Cell Hybrid Vehicle (FCHV). The main objective of this paper is to develop control strategies for a real vehicle which has been designed in order to demonstrate the feasibility of the use of hydrogen as an energy source for automotive applications. The hybrid vehicle is driven by an array of batteries and a bank of two fuel cells. The control system is designed and tested in order to perform all the operations related to the coordinated operation of the fuel cells, the intermediate electrical storage and the power train. Real-time power management is done in order to satisfy the power demand of the electric motor taking into account the state of charge of the batteries and the operating regime of the fuel cells. An embedded controller has been installed to manage the power distribution between the fuel-cell bank and the batteries and a supervisor has also been designed and tested. An important contribution to be noticed is the development of a power management strategy devised to split the power between both fuel cells. Besides, a low-level controller has been implemented to regulate temperature and air excess ratio in the cells. Some experimental results are presented in order to show the vehicle performance. The proposed hybrid vehicle, tested under real driving conditions, has demonstrated a high robustness and reliability.

This paper compares global and local model predictive control (MPC) algorithms in a traffic network controlled by intelligent transportation system (ITS) signals (ramp metering and variable speed limits). It will be shown that local techniques have a suboptimal behavior and that centralized techniques are difficult to implement in real time. To deal with this problem, a local MPC with only one communication cycle at each sampling time is proposed. This controller improves the local controller performance, and although it is suboptimal with regard to the centralized controller behavior, it can be implemented in real time.

Model predictive control (MPC) has developed considerably in the last decades both in industry and in academia. This success is due to the fact that MPC is perhaps the most general way of posing the control problem in the time domain. One of the main advantages of MPC is that model uncertainties can be taken explicitly into account and this allows for the consideration of faulty process behavior. The receding control strategy used in MPC can be extended to the case of system identification by parameter bounding and furthermore to determine if a model is consistent with the obtained data in a receding horizon manner and this implicitly allows for fault detection. The paper shows how concepts arising from the fault detection and fault-tolerant design methods can be incorporated in an MPC framework and the advantages that can be gained by using MPC in this context.

This paper presents a way of implementing a model-based predictive controller (MBPC) for mobile robot navigation using genetic algorithms to perform online nonlinear optimization. Obstacle avoidance is considered by the controller. Constraints on the maximum attainable speeds are also considered by the algorithm. Experimental results obtained when applying the controller to a TRC LABMATE mobile platform are given in the paper.

This paper presents an application of fuzzy logic control to the distributed collector field of a solar power plant. The major characteristic of a solar power plant is that the primary energy source, solar radiation, cannot be manipulated. Solar radiation varies throughout the day, causing changes in plant dynamics and strong perturbations in the process. A special subclass of fuzzy inference systems, the TP (triangular partition) and TPE (triangular partition with evenly spaced midpoints) systems, is used to obtain adequate control signals in the whole range of possible operating conditions. The fuzzy logic controller has been tested in the real plant and results obtained are shown. A comparison with other control approaches widely used in the plant is performed using a nonlinear computer model of the field.

In this paper, an adaptive controller for robot manipulators which uses neural networks is presented. The proposed control scheme is based on PD feedback plus a feedforward compensation of full robot dynamics. The feedforward signal is obtained by summing up the weighted outputs of a set of fixed multilayer neural nets. The controller is adaptive to robot dynamics and payload uncertainties. A stability analysis which takes into account neural network learning errors is included. Simulation results showing the feasibility and performance of the approach are given.>