Nonlinear model predictive control (MPC) and reinforcement learning (RL) are two powerful control strategies with complementary advantages. This work shows how actor-critic RL techniques can be leveraged to improve the performance of MPC. The RL critic is used as an approximation of the optimal value function, and an actor rollout provides an initial guess for the primal variables of the MPC. A parallel control architecture is proposed where each MPC instance is solved twice for different initial guesses. Besides the actor rollout initialization, a shifted initialization from the previous solution is used. The control actions from the lowest-cost trajectory are applied to the system at each time step. We provide some theoretical justification of the proposed algorithm by establishing that the discounted closed-loop cost is upper-bounded by the discounted closed-loop cost of the original RL actor plus an error term that depends on the (sub)optimality of the RL actor and the accuracy of the critic. These results do not require globally optimal solutions and indicate that larger horizons mitigate the effect of errors in the critic approximation. The proposed algorithm is intended for applications where standard methods to construct terminal costs or constraints for MPC are impractical. The approach is demonstrated in an illustrative toy example and an autonomous driving overtaking scenario.

This article proposes a control-based framework for aligning large language models (LLMs) by leveraging a control barrier function (CBF) to ensure user-desirable text generation. The presented framework applies the CBF safety filter to the predicted token generated from the baseline LLM to intervene in the generated text. The safety filter includes two significant advantages: this safety filter is an add-on type, allowing it to be used for alignment purposes without fine-tuning the baseline LLM, and if there is an evaluation model regarding the desired alignment, it can be directly applied to the filter design. The overall text-generation system is implemented with open-source language models, aiming to generate positive text.

Input-output linearization is a powerful analytical tool to support the design of controllers for vehicles characterized by nonholonomic constraints. Nonetheless, the presence of actuator dynamics and of significant delays, particularly in the steering actuation, can significantly hamper the applicability of this technique. This article proposes an extension to an input-output linearization approach presented in the literature, which allows for including the actuator dynamics into the linearization law, and to compensate for the effect of delays and nonidealities on the linearized system, introducing a further control action, designed with the advanced gain scheduling technique. As a result, a linear time-invariant decoupled system is obtained, which allows us to design the trajectory tracking controller using linear control tools. Simulation results, based on three different vehicle simulators, ranging from a single-track model with linear and nonlinear tire models to a dual-track model, are also presented. The performance and robustness of the inner controller, composed of the linearizing law and the advanced gain scheduling controller, are thoroughly analyzed and compared with the results obtained using the former input-output linearization available in the literature. Furthermore, the results of the application of the proposed control framework as part of a trajectory tracking controller show the efficacy of the proposal.

In this article, we present a novel control strategy for optimizing the management of reconfigurable batteries, particularly in electric vehicles (EVs) and energy storage systems. Our approach uniquely integrates control allocation (CA) with control barrier functions (CBF) to enhance operational efficiency while enforcing electrothermal safety constraints. Specifically, the proposed control scheme regulates the output voltage and ensures state-of-charge (SoC) balancing across battery modules, addressing the common weakest-module problem. Additionally, it dynamically adjusts power distribution to provide over-temperature protection, mitigating the risk of thermal runaway. Our feasibility analysis establishes the conditions under which the CBF-based safety constraints are satisfied, ensuring the existence of feasible control actions. Furthermore, we define comprehensive performance metrics to evaluate and compare the controller’s effectiveness. This includes CBF gain tuning based on feasibility results. We perform scalability analysis, demonstrating the controller’s effectiveness in systems with a large number of modules, which is critical for EV applications where computational efficiency is paramount. Finally, experimental validation on a small-scale prototype showcases the practical capability of our CBF-based optimal controller.

A novel topology-aware model predictive control (MPC) framework is presented for navigating off-road wheeled vehicles with extreme mobility on 3-D terrains. Prior studies have oversimplified the problem by either treating it solely as a path-planning problem or disregarding the terrain topology. To bridge this gap, this article presents a one-layer model predictive planning and control framework that considers terrain topology for off-road vehicles at operationally relevant speeds. The algorithm is tested in various simulated scenarios with comparisons to benchmarks. The proposed algorithm demonstrates superior performance by capturing the dynamic constraints, while the benchmarks either fail to complete tasks or exhibit inferior performance due to inaccurate representation of the dynamic limitations of the vehicle. Furthermore, the algorithm is tested for robustness to demonstrate its ability to handle uncertainties in terrain topology.

This article presents a closed-form variable-structure control strategy with dual avoidance regions design for collision-free waypoint navigation and trajectory tracking of car-like vehicles. New Bézier curve-based collision risk assessment functions are introduced to improve curvature adjustment for both cooperative and noncooperative avoidance pairs of vehicles. The variable-structure control guides the vehicle’s state toward a switching surface, stabilizing the risk function’s rate of change and reducing conservative avoidance behavior. Safety-corrected tracking control enhances performance in dense obstacle environments. Based on the Lyapunov approach, technical guarantees are formulated and proved for the system’s convergence and collision-free maneuvering. Two simulations illustrate the approach: one for point-to-point tracking compared with other approaches, and the other for cooperative vehicles navigating a dense obstacle field. Finally, an experiment validates the control law’s applicability and effectiveness.

Model predictive control (MPC) has demonstrated significant potential in improving energy efficiency in building climate control, outperforming traditional controllers commonly used in modern building management systems (BMSs). Among MPC variants, data-driven predictive control (DPC) offers the advantage of modeling building dynamics directly from data, thereby substantially reducing commissioning efforts. However, inevitable model uncertainties and measurement noise can result in comfort violations, even with dedicated MPC setups. This article introduces a disturbance-adaptive DPC (DAD-DPC) framework that ensures asymptotic satisfaction of predefined violation bounds without knowing the uncertainty and noise distributions. The framework employs a data-driven pipeline based on Willems’ Fundamental Lemma and conformal prediction for application in building climate control. The proposed DAD-DPC framework was validated through four building cases using the high-fidelity building optimization testing framework (BOPTEST) simulation platform and an occupied campus building, Polydome. DAD-DPC successfully regulated the average comfort violations to meet predefined bounds. Notably, the 5% violation DAD-DPC setup achieved 30.1%/11.2%/27.1%/20.5% energy savings compared to default controllers across four cases. These results demonstrate the framework’s effectiveness in balancing energy consumption and comfort violations, offering a practical solution for building climate control applications.