Being driving intrinsically a control task, control-oriented tools that capture how humans drive can promote trustworthy transfer of human capabilities to automated driving. This work presents an Inverse Reinforcement Learning (IRL) approach for capturing vehicle-following behavior from traffic data, where the control perspective intrinsic to driving is reproduced via a Lyapunov function acting as the value function of an optimal control problem. The idea is to describe vehicle-following as an optimal control problem with cost and constraints expressed and learned in terms of data. A distinguishing feature is to assume practical stability of the control task, instead of asymptotic stability as in standard approaches. This assumption, motivated by the non-asymptotic behavior exhibited by real-world data, is verified using the highD traffic dataset as case study: enforcing asymptotic stability may lead to infeasible learning. For a wide range of settings (linear or nonlinear vehicle-following dynamics, parametric or non-parametric value functions), the learning process of the proposed IRL boils down to a Quadratic Programming solvable using standard optimization routines. The approach is validated with tests in highD and CommonRoad, showing consistency even when changing the training data samples and the form of the value function.

PieceWise Affine (PWA) approximations for nonlinear functions have been extensively used for tractable, computationally efficient control of nonlinear systems. However, reaching a desired approximation accuracy without prior information about the behavior of the nonlinear systems remains a challenge in the function approximation and control literature. As the name suggests, PWA approximation aims at approximating a nonlinear function or system by dividing the domain into multiple subregions where the nonlinear function or dynamics is approximated locally by an affine function also called local mode. Without prior knowledge of the form of the nonlinearity, the required number of modes, the locations of the subregions, and the local approximations need to be optimized simultaneously, which becomes highly complex for large-scale systems with multi-dimensional nonlinear functions. This paper introduces a novel approach for PWA approximation of multi-dimensional nonlinear systems, utilizing a hinging hyperplane formalism for cut-based partitioning of the domain. The complexity of the PWA approximation is iteratively increased until reaching the desired accuracy level. Further, the tractable cut definitions allow for different forms of subregions, as well as the ability to impose continuity constraints on the PWA approximation. The methodology is explained via multiple examples and its performance is compared to two existing approaches through case studies, showcasing its efficacy.

This paper addresses the challenge of simultaneously compensating for state-dependent uncertainties and enforcing time-varying state constraints in Euler-Lagrange systems, a common requirement in robotics that remains underserved by existing control designs. A novel adaptive control framework is developed that combines an artificial time-delay-based uncertainty estimation strategy, also known as time-delay estimation, with a barrier Lyapunov function to enforce constraint-aware control design. Specifically, a state-dependent upper bound on the time-delay estimation approximation error is analytically formulated, and an adaptive law is constructed to estimate its parameters online, enabling real-time state-dependent uncertainty compensation without relying on prior model knowledge. To ensure constraint compliance, the barrier Lyapunov function-based controller enforces time-varying bounds on both position and velocity. The resulting architecture is provably stable via Lyapunov analysis. Experimental results on a five-degree-of-freedom robotic manipulator validate the framework's capability, compared with the state of the art, in maintaining strict adherence to safety-critical constraints under dynamic uncertainties.

The vast majority of autopilot architectures for rovers rely on a decoupled paradigm where the control algorithms in the loops composing the autopilot (typically, throttle control and steering control loops) are designed independently. We propose a novel autopilot paradigm in which the control algorithms are designed to handle not only possible couplings between loops, but also unknown structure of the couplings. Throttle and steering loops are intrinsically coupled when the vehicle needs to change direction, and it is hard to have a priori knowledge about the structure of the couplings. In the presented paradigm, adaptive laws are designed to compensate for the coupled and structurally unknown couplings. Stability analysis is performed in the Lyapunov framework, and hardware-in-the-loop experiments are presented using ArduPilot, a popular autopilot suite in the field.

The goal of model reference adaptive control (MRAC) is to ensure that the trajectories of an unknown dynamical system track those of a given reference model. This is done by means of a feedback controller that adaptively changes its gains using data collected online from the closed-loop system. One of the approaches to solve the MRAC problem is to impose conditions on the data that guarantee convergence of the gains to a solution of the so-called matching equations. In the literature, various extensions of the concept of persistent excitation have been proposed in an effort to weaken the conditions on the data required for this convergence.Despite these efforts, it is not well-understood what are the weakest possible data requirements ensuring convergence of MRAC. In this paper, we propose a new framework to study the MRAC problem, using the concept of data informativity. Our main contribution is to provide necessary and sufficient conditions for the asymptotic convergence of the adaptive gains to a solution of the matching equations. These necessary and sufficient conditions can be readily checked online as new data are generated by the closed-loop system. Our results reveal that existing excitation conditions impose stronger requirements on the collected data than required. Notably, the necessary and sufficient conditions provided in this paper are weaker than those for unique system identification.

This paper proposes an impulsive adaptive control (IAC) law for stabilization of uncertain linear systems. Available impulsive control results are based on impulsive fixed-gain control, where the impulsive control gain is tightly coupled with the system parameters and the impulsive frequency, thus limiting the capability to tackle uncertainty. Instead of a fixed gain, the proposed control law adaptively updates the gain at impulsive instants. The fact that adaptation occurs at discrete time instants can result in a more practical implementation as compared to IAC with continuous-time adaptation, i.e., requiring one to solve adaptive equations in a continuous fashion. Stability conditions for the proposed IAC protocol are established, showing less conservatism than impulsive fixed-gain control and more capability to handle uncertainties in the system and in the impulsive frequency. Numerical comparisons further verify the effectiveness of the proposed IAC scheme.

Hydrogen-enabled multi-energy systems (HMESs) have received much attention due to their low carbon emissions, high cross-sectoral/temporal/spatial flexibility, and black-start capability. However, their management under extreme weather events is particularly challenging due to the interacting multi-timescale dynamics of electrical-thermal-hydrogen subsystems, spatiotemporally coupled constraints, and large-scale mixedinteger decision variables. In this paper, we investigate a risk-aware economic scheduling problem for building-scale HMESs and aim to address all the above challenges. Specifically, by taking multi-timescale energy-coupled dynamics, extreme weatherinduced source-load impacts, priority-based load shedding, and mobile emergency energy resources (e.g., electric vehicles and hydrogen fuel cell vehicles) into consideration, we formulate a risk-aware economic scheduling problem with the aim of minimizing the weighted sum of expected operation cost and conditional value-at-risk by optimizing the ex-ante rental quantities of mobile emergency energy resources. To solve this large-scale mixed-integer programming problem, we propose an algorithm that combines a farthest-first traversal-based clustering method with an improved quantum-behaved particle swarm optimizer (QBPSO)-based decomposition method. Simulation results indicate the effectiveness of the proposed algorithm.

The sudden appearance of a static obstacle on the road, i.e. the moose test, is a well-known emergency scenario in collision avoidance for automated driving. Model Predictive Control (MPC) has long been employed for planning and control of automated vehicles in the state of the art. However, real-time implementation of automated collision avoidance in emergency scenarios such as the moose test remains unaddressed due to the high computational demand of MPC for evasive action in such hazardous scenarios. This paper offers new insights into real-time collision avoidance via the experimental imple- mentation of MPC for motion planning after a sudden and unexpected appearance of a static obstacle. As the state-of-the-art nonlinear MPC shows limited capability to provide an acceptable solution in real-time, we propose a human-like feed-forward planner to assist when the MPC optimization problem is either infeasible or unable to find a suitable solution due to the poor quality of its initial guess. We introduce the concept of maximum steering maneuver to design the feed-forward planner and mimic a human-like reaction after detecting the static obstacle on the road. Real-life experiments are conducted across various speeds and level of emergency using FPEV2-Kanon electric vehicle. Moreover, we demonstrate the effectiveness of our planning strategy via comparison with the state-of- the-art MPC motion planner.

Multi-modal vehicle trajectory prediction is crucial for autonomous driving in dynamic environments. Despite the significant progress in the field, the uncertainty and heterogeneity caused by the diversity in driving intentions and driving scenes still present major challenges to multi-modal prediction. Existing query-based prediction paradigms take into account the social context arising from the driving scene, but neglect priors regarding vehicle intentions that the historical trajectory may contain. We propose a Social-aware Water-wave Information Transmission Recurrent Acceleration Network, abbreviated as S-WITRAN, based on decoupling multi-modal prediction into an ego-aware and a social-aware learning stage. The ego-aware stage aims to relax the constraints from the driving scene to explore a diversity of future trajectory candidates. The social-aware stage aligns the candidates with respect to the social context arising from the driving scene. The information transmission is designed to extract from the vehicle’s historical trajectory priors about its possible intentions and dynamic states, which are integrated to form a multi-modal set of trajectories. Extensive experiments on the NGSIM, highD, and Argoverse2 datasets, as well as a benchmark evaluation in UniTraj, demonstrate the superior performance of the proposed model in both map-free and map-based datasets.

In this study, a hybrid air-sea swarm control methodology is introduced to achieve resilience against denial-of-service (DoS) attacks, where such attacks intermittently block communication links to disrupt coordination. We start by establishing a swarm system comprising uncrewed aerial vehicles (UAVs) and uncrewed surface vehicles (USVs) with distinct dynamic properties and degrees of connectivity. Unlike existing studies with homogeneous communication links, we consider heterogeneous (hybrid) UAV-to-UAV, USV-to-USV, and UAV-to-USV communication links. A cooperative-competitive interaction strategy is studied, where the swarm is divided into two clusters based on vehicle type. To face cyber vulnerabilities, an event-triggering security-oriented control law is designed, which is proven to counteract attacks while balancing communication and control resource usage. Furthermore, the relation between communication weights and system resilience to DoS attacks is analyzed. Numerical simulations demonstrate the efficacy and resilience of the proposed approach.

This work proposes a guaranteed collision avoidance framework for Cooperative Adaptive Cruise Control of a vehicular platoon characterized by unidirectional communication and heterogeneous parameters. In the proposed framework, the actual (heterogeneous) platoon is made to converge to a reference (homogeneous) platoon via adaptive laws designed using of set-theoretic model reference adaptive control. Yet, in contrast to the state-of-art that is based on ensuring collision avoidance on the reference platoon dynamics only, the approach we propose can ensure collision avoidance on the actual platoon dynamics. This result is possible thanks to the introduction of a novel concept of virtual platoon, only used for analysis, but that does not interact with the actual platoon. The stability and convergence properties of the proposed framework are established using Lyapunov-based analysis in conjunction with the aforementioned virtual platoon concept. Simulation results further demonstrate the effectiveness of the proposed framework.

Embedding adaptation capabilities in autopilots is crucial to handle a wide range of uncertain conditions arising from environmental disturbances, varying loads, and non nominal regimes. In this work, we release AdArduPilot+, the enhancement of an adaptive autopilot fully compatible with the ArduPilot suite. Key features of AdArduPilot+ are its adaptive laws with the ability to handle uncertainties with either linear in-parameter (LIP) or nonlinear-in-parameter (NLIP) nature. This hybrid adaptation allows the autopilot to negotiate between structured and unstructured uncertainty: NLIP adaptation uses an upper-bounding mechanism that does not require structural knowledge of the uncertainty, while LIP adaptation can reduce such upper bound by compensating the linear portion of the uncertainty. Stability analysis of AdArduPilot+ does not rely on strong structural assumptions, allowing the autopilot to flexibly cope with different vehicle structures. Experiments on fixed-wing and rotary-wing vehicles validate the flexibility of the proposed solution and show that the hybrid adaptation of AdArduPilot+ outperforms the traditional autopilot and two adaptive variants.

The digital implementation of any energy management program faces discrepancies between continuously evolving physical variables (energy, temperature) and discretely executed automation routines: discrepancies are also inevitable between the ground-truth system and the control-oriented dynamics used to make the energy management real-time tractable. This work provides a theoretical and numerical framework to study robustness of digitally implemented energy management to such discrepancies. The theoretical challenge solved by our robustness framework lies in analytically handling rather general and possibly time-varying energy management problems operating digitally under a wide range of discrepancies. The analytical existence of robustness margins is confirmed via a numerical energy benchmark we introduce and for which control-oriented dynamics are derived. Numerical comparisons under established digitization strategies show that, despite modeling errors ranging in 6-52% and discretization errors ranging in 2-23%, the performance of model-based energy management (implemented via linear quadratic control) is more robust than rule-based or proportional-integral-derivative energy management. Such results are consistently verified under different sampling times, automation and grid-interactive demand response scenarios.

Although the benefits of platoons of heavy-duty vehicles in terms of air drag reduction and fuel/energy savings are well recognized, uncertainties and heterogeneities in the vehicles composing the platoon can degrade such ideal benefits and even compromise inter-vehicle safety. This work proposes a system-wide platooning strategy for heavy-duty vehicles that accounts for uncertainty, heterogeneity, air drag reduction and safety. The strategy is based on two key principles: adaptation and self-organization. Adaptation aims at canceling uncertainty by making the vehicles homogenize to desirable reference dynamics. Self-organization aims at adjusting the inter-vehicle headway to increase fuel/energy savings via drag reduction, while accounting for (un)safety due to communication/actuation delay. The two principles are complementary: adaptation without self-organization would fail to attain substantial drag reduction, while self-organization without adaptation would result in reduced safety due to the different vehicle responses to acceleration commands. A mathematical formulation for these principles is given, and their effectiveness is validated in a SUMO-Veins co-simulation environment and with real-world data from the Heavy Heavy-Duty Diesel Truck (HHDDT) driving cycle of the California Air Resources Board.

The vast majority of protocols for connected automated vehicles are based on status sharing, i.e., communication of the current vehicle state among neighboring vehicles. Only recently the idea of intent sharing has been put forward, where not only the current state, but also the vehicle intention in the near future can be communicated. In the context of Cooperative Adaptive Cruise Control (CACC), this work provides a system-theoretic framework for intent sharing through the lens of output regulation. We present analytical results showing two fundamental aspects of CACC with intent sharing: a) when vehicle-to-vehicle communication is reliable, intent sharing provides no benefits over status sharing, as both sharing paradigms result in the same protocol; b) intent sharing becomes beneficial when vehicle-to-vehicle communication is unreliable, in which case the latest communicated intent can be used to reconstruct the missing information of the neighboring vehicle in the near future. Together with theoretical analysis, numerical validations with synthetic and real-world data are provided, where the benefits of CACC with the proposed implementation of intent sharing are shown against several state-of-the-art CACC protocols.

We address the problem of controlling a hierarchical system composed of high-level and low-level loops, subject to both matched and unmatched uncertainties with a priori unknown structure. To tackle this challenge, we employ an adaptive observer in the low-level loop to estimate the uncertainties and to make the controlled system behave as close as possible to the nominal model used to design the high-level control. A stability analysis of the hierarchical control architecture is presented. Comparative experiments on a flight control numerical benchmark demonstrate that the proposed approach outperforms several state-of-the-art methods, including those based on fixed control gains or adaptive control gains.

Most literature in the field of vehicular platooning considers control-oriented lagged vehicle dynamics where the vehicle acceleration tracks, with some lag, a control command given in terms of desired acceleration/deceleration (hereafter simply referred to as desired acceleration). However, the most typical control command in standard vehicles is the desired throttle/braking (hereafter simply referred to as desired throttle). This work proposes an interface design that allows relating the desired throttle to the desired acceleration, enabling the use of those control-oriented vehicle models in the platooning literature. A challenge in the design of such an interface is that the relation between the desired throttle and the desired acceleration is dynamic and, furthermore, the design must account for the lagged dynamics arising from the time constant of the vehicle powertrain/driveline trying to track the desired command. This challenge adds to the fact that several vehicle parameters, including the powertrain/driveline time constant, are unknown in practice, requiring an adaptive design of the interface able to deal with uncertainties. We initially propose a single-vehicle adaptive design, and then extend it at the platooning level. Theoretical stability analysis is provided, and robustness of the proposed ideas is demonstrated via software-in-the-loop tests using an open-source autopilot suite embedded with the proposed adaptive interface.

This paper provides a data-driven design guaranteeing parameter convergence in model reference adaptive control (MRAC) when the to-be-controlled system is subject to process noise. In the context of MRAC, parameter convergence refers to ensuring convergence of the adaptive gains to a solution of the matching equations, or to an approximate solution when noise is present. In classical MRAC, even small noise may induce parameter drift, thus lacking robustness to noise. Meanwhile, existing robust MRAC methods cannot ensure parameter convergence without imposing excitation conditions on data. A key feature of the proposed framework is to ensure convergence of the adaptive gains to an approximate solution of the matching equations without relying on persistently exciting signals. Furthermore, the matching error can be explicitly characterized as a function of the noise. This explicit characterization allows to establish a necessary and sufficient condition on the noise characteristics under which the limit closed-loop system matrix is Hurwitz. In the noise-free case, the proposed framework results in exact parameter convergence. Notably, as compared to existing methods achieving exact parameter convergence in the noise-free case, the condition on data in the proposed framework is weaker.

One of the major challenges in deploying demand-responsive energy management programs is the design of control-oriented models that explain how the demand is affected by shifting commands (e.g., shifts in the set points of lights, heating, ventilation and air conditioning (HVAC) units). This work proposes an averaging technique for control-oriented demand modeling: the key aspect resides in determining the average thermostatic phases emerging from coupled continuous-discrete demand dynamics. We analytically study stability, optimality, and robustness of averaging-based energy management, and we benchmark its effectiveness against state-of-the-art learning-based energy management. The proposed approach exhibits 3-4 times improved demand-responsive tracking, without compromising user's flexibility in terms of set points and temperature deviations. Notably, the thermostatic phases corrupt the effectiveness of learning-based energy management, but not of the proposed averaging-based energy management. Note to Practitioners-Thermostatic ON/OFF phases triggered by set point shifts complicate the demand modeling by creating coupled continuous-discrete dynamics. In the proposed approach, averaging the phases directly supports the calculation of the demand as a function of the set points. Shifting commands can thus be obtained dynamically as system conditions change. Outperforming state-of-the-art learning-based energy management confirms the validity of such averaging technique in tackling complex time-varying demand dynamics.

Spacing policies defined in terms of a constant time headway between adjacent vehicles are traditional and ubiquitous in the vehicular platooning literature. This paper presents a new platooning framework where the traditional constant time headway is generalised to a possibly non-constant (and thus more flexible) time headway. We first consider the ideal scenario where the vehicle dynamics are known, and give design conditions for making the proposed generalised time headway satisfy the same desirable properties reported in the literature for the traditional constant time headway, namely, disturbance decoupling, string stability and collision avoidance. Then, we turn the attention to the case of uncertain vehicle dynamics, solving the vehicular platooning problem via two ingredients: a reference model designed to satisfy the aforementioned desirable properties for nominal vehicle dynamics, and adaptive laws designed to make the actual dynamics of each vehicle in the platoon converge to the nominal dynamics. Lyapunov-based stability analysis is conducted to derive two types of adaptation laws: direct and indirect. Both designs allow the closed-loop system to converge to the desirable properties of the reference model. Numerical simulations are conducted to validate the newly proposed generalised spacing policy and its adaptive versions: the simulations illustrate the flexibility of non-constant time headway over a constant one, especially showing that the proposed time headway is capable of modulating itself during acceleration/deceleration phases.