Solar sail momentum management with mass translation and reflectivity devices using predictive control

Studies in humans and monkeys have shown that action observation activates not only cortical areas of the mirror neuron system (MNS) but also the cerebellum and other subcortical structures. Cortico-cerebellar circuits are proposed to be involved in the predictive control and simulation of goal-directed observed actions. However, it remains unclear whether cerebellar projections originating from visuo-motor sectors are partially segregated from those starting from purely motor sectors. To address this issue, sixteen healthy participants underwent an fMRI study in which they were required to observe and execute grasping actions. The study allowed to identify cerebellar and thalamic regions predominantly involved in motor execution, as well as regions activated during both observation and execution. Probabilistic tractography and effective connectivity analyses were then used to characterize the projections and functional interactions of these sectors. The cerebellar lobules I-V, dorsal dentate nucleus (DN) and ventrolateral-ventral anterior thalamic nuclei (VL-VA) were mainly active during execution, whereas cerebellar lobule VI, ventral DN, red nucleus (RN) and ventroposterolateral thalamic nucleus (VPL) showed shared activation during observation and execution. At the cortical level, both observation and execution engaged the ventral premotor cortex (PMv) and the inferior parietal lobule (IPL). Tractography revealed that dorsal DN tracts project to RN and VL-VA, terminating in rostral IPL and ventral PMv, while ventral DN projections target RN and VPL, terminating more caudally in IPL and more dorsally in PMv. Effective connectivity analyses showed that execution was associated with increased coupling between DN, RN, VL-VA, and IPL-PMv, while both observation and execution were characterized by enhanced connectivity between DN, RN, VPL, and IPL-PMv. Overall, these findings indicate that cerebellar projections to thalamic and cortical regions involved in action observation and execution are partially segregated from purely motor projections, supporting a cerebellar role in motor simulation during observation and execution of goal-directed actions.

Optimal energy management of buildings using neural network-based thermal prediction and economic model predictive control

In the sensorless control of surface-mounted permanent magnet synchronous motors (SPMSMs), the model reference adaptive system (MRAS) is known for its favorable dynamic performance. However, the conventional MRAS method, which estimates the rotor position by integrating the speed estimation results, inherently results in a relatively slow convergence rate of position errors during dynamic control. To address this issue, a dual MRAS sensorless control method based on model predictive current control (MPCC) is proposed in this paper. A novel dual MRAS scheme for rotor speed and position error estimation is constructed, leveraging two components of the total electromotive force (EMF) in the synchronous rotating frame through distinct estimation and prediction models. Small-signal analysis reveals that the position error equation transitions from a second order system in the traditional method to a zero-order system in the proposed method, thereby enabling faster response and reduced dynamic position estimation errors. Furthermore, by incorporating MPCC into the current loop, the proposed method achieves a closed-loop prediction, thereby enhancing estimation accuracy and the dynamic response performance. Comprehensive experimental validations are conducted to evaluate the performance of the proposed dual MRAS method based on MPCC (D-MRAS-MPCC)methodology. The results demonstrate that the proposed method exhibits superior dynamic control performance compared to the conventional MRAS method.

Model predictive torque control for SRM with switching states restructuration and online torque compensation

Harmonic currents inherently exist in a dual three-phase permanent magnet synchronous motor (DTP-PMSM) due to various disturbances, such as inverter nonlinearity and motor asymmetry. This letter proposes a data-driven deadbeat predictive harmonic current control (DD-DPHCC) method aimed at suppressing harmonic currents. First, a data-driven model of the harmonic currents for a DTP-PMSM is developed using dynamic linearization theory. This model exclusively relies on system input and output data to facilitate real-time updates, thereby eliminating the necessity for motor parameters. Then, a deadbeat control algorithm is designed to directly calculate the harmonic control voltage based on the sampled harmonic currents and the data-driven model, ensuring precise harmonic current control performance. Compared to conventional model-based DPHCC methods, the proposed DD-DPHCC exhibits enhanced robustness against motor parameter mismatches. Finally, experimental results substantiate the superior performance of the proposed DD-DPHCC method in suppressing harmonic currents.

Event-triggered sliding mode predictive control for piezoelectric-actuated hybrid vibration isolators subject to hysteresis and time delays

The control of large buildings encounters challenges in computational efficiency due to their size and nonlinear components. To address these issues, this paper proposes a Piecewise Affine (PWA)-based distributed scheme for Model Predictive Control (MPC) that optimizes energy and comfort through PWA-based quadratic programming. We utilize the Alternating Direction Method of Multipliers (ADMM) for effective decomposition and apply the PWA technique to handle the nonlinear components. To solve the resulting large-scale nonconvex problems, the paper introduces a convex ADMM algorithm that transforms the nonconvex problem into a series of smaller convex problems, significantly enhancing computational efficiency. Furthermore, we demonstrate that the convex ADMM algorithm converges to a local optimum of the original problem. A case study involving 36 zones validates the effectiveness of the proposed method. Our proposed method reduces execution time by 86% compared to the centralized version.

Lyapunov-based model predictive control for path-following of an autonomous underwater vehicle using line-of-sight guidance

Integrated multi-objective model predictive control framework for cascaded open-channel systems with multi-source lateral inflows

Design and application of mixed-integer nonlinear model predictive control in residential buildings

Sliding-mode predictive control for constrained hovercraft: A short-horizon framework with global stability and external disturbances

Research on an improved model predictive control–based control strategy for a two-stage cascaded hydropower stations with a regulating reservoir

Water level predictive control in cascaded canals based on the variation pattern analysis of integrator-delay model parameters

A maximum power point tracking control for wind energy conversion systems using regularized data-enabled predictive control

Dynamic adaptive model predictive control for prosumers-based energy communities

Computationally efficient smart building energy management via deep reinforcement learning-enhanced model predictive control

Adaptive nonlinear model predictive control of monoclonal antibody glycosylation in CHO cell culture

Tiny learning-based MPC for multirotors: Solver-aware learning for efficient embedded predictive control

AI-driven real-time monitoring and predictive control system for yoghurt fermentation