Fuel cell electric vehicles (FCEVs) offer a sustainable transportation solution by combining high efficiency with zero local emissions. However, the complex dynamics of air-path management and the demands of real-time energy control pose significant challenges. This work proposes a non-linear model predictive control (NLMPC) strategy leveraging singular value decomposition (SVD) to obtain a reduced-order model from a detailed seven-state physical representation, including battery state-of-charge (SOC) and air-path dynamics. By preserving the system’s dominant dynamics, the reduced model allows for computationally efficient control within a CasADi-based MPC framework. The proposed approach is evaluated across multiple reduction levels, analyzing trade-offs between model complexity, computational cost, and hydrogen consumption. Simulations demonstrate fuel savings of up to 6.4%, with up to 3.3% achieved under the selected real-time feasibility threshold. Experimental data were used to calibrate and validate the high-fidelity plant model, and the control strategy was assessed in closed-loop simulations under realistic driving cycles.

The accelerated adoption of electric vehicles (EVs), driven by a growing global market and sustainability laws, has magnified the importance of Battery Thermal Management Systems (BTMS). The efficacy of these systems, however, is challenged by real-world variables, including traffic patterns, ambient conditions, and driver behavior. This work presents a predictive thermal management strategy that integrates driving and charging phases to minimize energy consumption over a complete trip. Leveraging the vehicle route information to anticipate charging stops and the State of Charge (SOC) allows to proactively adapt the control strategy to the current situation. When a charging event is predicted, the controller preconditions the system and transitions to a charging-optimized strategy that manages the thermal stresses associated with fast charging. The strategy was validated experimentally using a 3.7 kWh battery pack, demonstrating a reduction in BTMS energy consumption compared to a conventional rule-based controller across different drive cycles.

This paper aims to extend the potential of Adaptive Cruise Control (ACC) technology by forecasting the behaviour of the traffic using physical models. A preceding vehicle prediction algorithm is proposed estimating the future kinematics of the vehicles with an extended car-following model. The prediction is tested with two cruise control frameworks: An anticipative car-following (ACF) control where the prediction is used to anticipate the drivers reaction and an economical predictive cruise control algorithm, named as eco-PCC, where the estimation of the preceding vehicle is added as a constraint to an optimal control problem (OCP) to minimise the fuel consumption over a receding horizon. The algorithms are tested in SUMO with a validated simulation environment of the city of Bologna. Results compare the benefits of the ACF approach and the eco-PCC algorithm with a baseline ACC without predictions. Several prediction horizons have been tested, highlighting the trade-off between prediction accuracy and the energy improvements. Experiments in an engine testbed presented benefits up to 8.03% in fuel consumption for the ACF control and up to 24.14% for the eco-PCC algorithm when compared to the baseline ACC without prediction.

Hydrogen enrichment of compressed natural gas (HCNG) was investigated in a spark-ignition engine using an experimental framework to examine the combined effects of blend composition, load, knock limitation, and lean operability. The study comprised a baseline CNG spark sweep, stoichiometric hydrogen-fraction sweeps from 19.5 to 100 vol% at 3 bar, 6 bar, and full load, and lean-limit exploration. The baseline sweep assessed spark-timing effects on efficiency, peak pressure, and NOx emissions, while the parametric tests were conducted at CA50 approximate to 8 degrees aTDCf. Under stoichiometric operation, hydrogen shortened flame development by up to 73.1% and combustion duration by up to 72.9%. Carbon-based emissions decreased, whereas pressure-rise rates and knock propensity increased; full load became knock-limited above 68% H2. In lean sweeps, hydrogen extended the stable lambda range, and the results show that the most favorable HCNG blend depends on operating condition, supporting variable dual-fuel metering.

This paper assesses the efficiency limits of light-duty vehicle propulsion systems based on reciprocating internal combustion engines (ICE) in the current state of the art and in the next five-year horizon, considering their combination with technologies such as electric turbocharging and hybridization, while excluding plug-in hybrid configurations so that fuel remains the primary onboard energy source. A systematic methodology is applied to evaluate the influence of key variables—heat transfer, air–fuel ratio, and compression ratio—on engine performance, integrating these variations into a simulation model to capture their interactions and effects. The resulting parametric study enables the generation of new engine maps that exploit synergies between parameters and enhance the prediction of engine behaviour across different operating conditions, forming the basis for assessing potential advancements in hybrid powertrain architectures. These maps are then used to define performance expectations for hybrid vehicles, identifying optimal parameter combinations to guide future technology development and improve efficiency in hybrid powertrain design. The proposed powertrain architectures are integrated into a representative vehicle model, considering two vehicle typologies: a compact passenger car and a sport utility vehicle (SUV). To quantify the potential fuel-consumption benefits, an intelligent energy-management algorithm is implemented to supervise and optimize system operation over a WLTC driving cycle. The results indicate that the proposed configurations can achieve fuel-consumption reductions exceeding 20%, demonstrating the effectiveness of both the powertrain designs and the control strategies. Overall, the findings highlight the significant efficiency potential of advanced ICE-based propulsion systems when combined with near-term technologies such as electric boosting and hybridization, confirming the viability of these improvements and providing a robust basis for future hybrid vehicle development focused on maximizing energy efficiency in transportation.

Dynamic Programming (DP) is often used to compute the optimal energy management in fuel cell vehicles (FCV) during a priori known driving cycles to benchmark different technologies or provide insight into suitable control strategies to be applied online. Due to the curse of dimensionality, using DP usually involves the use of simplified models that apply the quasi-steady approach for FC modelling, employing a map that provides the net FC power for a given current demand. While electro-chemistry processes inside the FC are much faster than the driving cycle dynamics, the response of the air-path, specially if turbocharging is used may make the quasi-steady hypothesis too optimistic. In this work, a state-of-the-art FCV model with seven states, adapted from the literature, has been calibrated using experimental data for the FC. The performance of using the quasi-steady approach for DP optimization has been assessed, leading to errors in H2 consumption above 13% for the considered cycles. Then, a model order reduction technique based on the Singular Value Decomposition (SVD) is applied to enable the use of DP on a simplified model including FC dynamics with positive results reducing the gap between the global and simplified model to levels lower than 10% and providing benefits in H2 consumption of 3.7 and 1.3% in WLTC and RDE cycles, respectively. Copyright (c) 2025 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

This paper deals with minimisation of fuel consumption under real driving conditions using a vehicle speed advisor. The aim is to explore the potential of speed profile optimisation in real driving conditions while assessing the suitability of an application which recommends the driver the optimal vehicle speed sequence that minimises the fuel consumption on a particular route. The speed advisor is based on solving the Optimal Control problem of covering a particular route with minimum fuel consumption with a defined time constraint. The approach presented was applied to and implemented on a real passenger vehicle to obtain a trade-off between fuel consumption and travel time for several trips on the route. Experimental results are presented with and without advisory. With speed advisor, the results approach the pareto front with lesser dispersion. On the other hand, without advisory, the dispersion is higher and largely above the pareto front.

This paper introduces a supervisory controller designed for Plug-in Hybrid Electric Vehicles (PHEVs) to minimize total energy consumption and tailpipe NOx emissions while adhering to Zero Emission Zone (ZEZ) constraints. The case study exploits historical driving cycle data from an urban bus route in Zurich to analyze trip correlations, where a ZEZ restriction was added to assess the vehicle performance under such conditions. A PHEV bus model was built, integrating the powertrain and After-Treatment System (ATS), where an electric heater is included to mitigate NOx emissions. The supervisory controller is tasked with determining the optimal power split and the electric heater power to ensure adherence to feasible operating conditions along the route. Historical driving cycle data analysis demonstrates that the speed profiles along the selected route exhibit similarities. This observation is leveraged by a Dynamic Programming (DP) optimization, where an arbitrary bus trip is employed to generate a cost-to-go matrix. Then, the resulting cos-to-go matrix was indexed by the position in the route and is used on the real-time controller by applying the one-step look-ahead roll-out algorithm. Simulation results demonstrate the controller effectiveness in addressing ZEZ restrictions, presenting a trade-off between energy consumption and tailpipe NOx emissions. A benchmark was carried out, comparing the results obtained with the DP solution as the baseline, assuming perfect knowledge of driving cycle disturbances, revealing a 7% increase in tailpipe NOx emissions and a 1.3% increase in fuel consumption compared to the theoretical minimum and fulfilling the ZEZ restrictions in all the cases.

This paper analyzes the control strategy for urban battery-swapping stations by optimizing the charging policy based on real-time battery demand and the time required for a full charge. The energy stored in available batteries serves as an electricity buffer, allowing energy to be drawn from the grid when costs or equivalent CO2 emissions are low. An optimized charging policy is derived using dynamic programming (DP), assuming average battery demand and accounting for both the costs and emissions associated with electricity consumption. The proposed algorithm uses a prediction of the expected traffic in the area as well as the expected cost of electricity on the net. Battery tests were conducted to assess charging time variability, and traffic density measurements were collected in the city of Valencia across multiple days to provide a realistic scenario, while real-time data of the electricity cost is integrated into the control proposal. The results show that incorporating traffic and electricity price forecasts into the control algorithm can reduce electricity costs by up to 11% and decrease associated CO2 emissions by more than 26%.

The present paper optimizes the driving profile for an electric vehicle, by using an optimal control formulation and route scheduling. The slope, traffic lights timing and the trajectory limitations, i.e., speed limitations and safety speed in turns, are computed along the route for minimising the energy consumption with consideration of the time. The work also analyses how unexpected traffic conditions caused by external drivers might influence the final performance of the controller. The paper uses real experimental data in several routes between Bath and the University of Bath from a HondaE electrical vehicle and reproduces several traffic conditions by using SUMO to validate the proposed methodology and determine its limitations under traffic disturbances. Copyright (c) 2025 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

Industry 4.0, leveraging tools like AI and the massive generation of data, is driving a paradigm shift in maintenance management. Specifically, in the realm of Artificial Intelligence (AI), traditionally “black box” models are now being unveiled through explainable AI techniques, which provide insights into model decision-making processes. This study addresses the underutilization of these techniques alongside On-Board Diagnostics data by maintenance management teams in urban bus fleets for addressing key issues affecting vehicle reliability and maintenance needs. In the context of urban bus fleets, diesel particulate filter regeneration processes frequently operate under suboptimal conditions, accelerating engine oil degradation and increasing maintenance costs. Due to limited documentation on the control system of the filter, the maintenance team faces obstacles in proposing solutions based on a comprehensive understanding of the system’s behavior and control logic. The objective of this study is to analyze and predict the various states during the diesel particulate filter regeneration process using Machine Learning and explainable artificial intelligence techniques. The insights obtained aim to provide the maintenance team with a deeper understanding of the filter’s control logic, enabling them to develop proposals grounded in a comprehensive understanding of the system. This study employs a combination of traditional Machine Learning models, including XGBoost, LightGBM, Random Forest, and Support Vector Machine. The target variable, representing three possible regeneration states, was transformed using a one-vs-rest approach, resulting in three binary classification tasks where each target state was individually classified against all other states. Additionally, explainable AI techniques such as Shapley Additive Explanations, Partial Dependence Plots, and Individual Conditional Expectation were applied to interpret and visualize the conditions influencing each regeneration state. The results successfully associate two states with specific operating conditions and establish operational thresholds for key variables, offering practical guidelines for optimizing the regeneration process.

Industry 4.0, leveraging tools like AI and the massive generation of data, is driving a paradigm shift in maintenance management. Specifically, in the realm of Artificial Intelligence (AI), traditionally “black box” models are now being unveiled through explainable AI techniques, which provide insights into model decision-making processes. This study addresses the underutilization of these techniques alongside On-Board Diagnostics data by maintenance management teams in urban bus fleets for addressing key issues affecting vehicle reliability and maintenance needs. In the context of urban bus fleets, diesel particulate filter regeneration processes frequently operate under suboptimal conditions, accelerating engine oil degradation and increasing maintenance costs. Due to limited documentation on the control system of the filter, the maintenance team faces obstacles in proposing solutions based on a comprehensive understanding of the system’s behavior and control logic. The objective of this study is to analyze and predict the various states during the diesel particulate filter regeneration process using Machine Learning and explainable artificial intelligence techniques. The insights obtained aim to provide the maintenance team with a deeper understanding of the filter’s control logic, enabling them to develop proposals grounded in a comprehensive understanding of the system. This study employs a combination of traditional Machine Learning models, including XGBoost, LightGBM, Random Forest, and Support Vector Machine. The target variable, representing three possible regeneration states, was transformed using a one-vs-rest approach, resulting in three binary classification tasks where each target state was individually classified against all other states. Additionally, explainable AI techniques such as Shapley Additive Explanations, Partial Dependence Plots, and Individual Conditional Expectation were applied to interpret and visualize the conditions influencing each regeneration state. The results successfully associate two states with specific operating conditions and establish operational thresholds for key variables, offering practical guidelines for optimizing the regeneration process.

Driving advisory usually require a-priori information about the route, which can be provided by context awareness technologies, such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity. This study presents a cooperative driving algorithm by formulating an optimal control problem (OCP) and solving it with a non-linear Model Predictive Control (MPC) approach with two steps. First, the OCP is solved by Dynamic Programming (DP), providing the cost-to-go (CTG) from any pair time-state to the end of the problem. The CTG can be used as a terminal cost for the MPC to be solved recursively during the route considering traffic as a disturbance in a short horizon. Secondly, the optimization is solved iteratively: at any time-step the first vehicle is solved assuming that the rest follow the traffic prediction algorithm. The process continues sequentially until all the vehicles have been solved, while the traffic prediction is replaced by the optimal trajectories of the vehicles that have been previously computed. Different prediction horizons, mixed traffic levels and the scheduling policy to solve the different vehicles are evaluated. Results suggest that the cooperative driving algorithm can reduce the fleet fuel consumption by up to 35%, while covering the same route in a similar time.

Traffic jams are one of the main causes of city pollution and significantly impact the economic cost of transportation. Context awareness by the traffic players may be key to improving the current control strategies and optimising traffic flow. This study investigates the effect of information availability through vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity in an urban real-driving route. An optimal control problem (OCP) is formulated to create speed advisory profiles, and it is solved using dynamic programming (DP) to provide the global optimal solution. Experimental engine tests have been used to characterise the fuel consumption and emissions of the engine, while traffic sensors around the city of Valencia have been used to reproduce realistic urban mobility using the traffic simulation software SUMO. The paper quantifies the impact of traffic information on vehicle fuel consumption and emissions. Under normal traffic conditions and assuming total access to the traffic information, the DP algorithm can reduce almost 60% on average the fuel consumption compared to normal driving behaviour provided by the default car-following model of SUMO.

The rapid adoption of Battery Electric Vehicles (BEVs) has been driven by growing environmental awareness and advancements in energy storage technologies. However, lithiumion cells, central to BEVs, are highly sensitive to temperature variations, requiring effective thermal management to prevent degradation, ensure safety, and optimize performance. This work presents a novel method to replicate the thermal behaviour of a battery on liquid cooling systems using standard system components. By combining a virtual battery model with a physical system, the thermal behaviour of a real battery pack is accurately reproduced. This cost-effective, safety-compliant approach enhances the development of efficient thermal management systems for BEVs. The Hardware In the Loop (HIL) platform was developed with a PXI from National Instruments and a solid-state resistance of 1 kW, while a 4 kWh battery pack prototype refrigerated with a cold plate was used for validation. Copyright (c) 2025 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

As regulations on pollutant emissions rapidly advance and the demand for sustainable mobility grows, the necessity for innovative technological solutions becomes crucial. To address these challenges, this research focuses on the development and application of a control-oriented model for plug-in hybrid electric vehicles (PHEVs), aimed at minimizing fuel consumption and NOx emissions while respecting operational constraints imposed during the vehicle’s operation. Accordingly, the model developed integrates the powertrain and the after-treatment system based on non-linear model predictive control (NLMPC) framework, strategically modulating the power distribution between the internal combustion engine (ICE) and the electric motor (EM), along with the ammonia injection strategy for effective NOx abatement and fuel savings. To overcome the finite horizon limitations of NLMPC, an offline dynamic programming (DP) was embedded, improving predictive capabilities through a cost-to-go matrix that reflects optimal control actions under specific conditions. This hybrid approach combines the global optimization of DP with the real-time flexibility of NLMPC, allowing dynamic adjustments to vehicle operation in response to real-time data and future scenarios. The applicability of the proposed strategy is demonstrated in routes containing a zero-emission zone and vehicles with different battery sizes, underlining its adaptability to complex driving conditions and distinct vehicle designs, thereby demonstrating its potential for significant contributions to sustainable mobility solutions.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

The development of electrified vehicles is a promising step toward energy savings, emissions reduction, environmental protection, and more sustainable economic growth. In the case of hybrid electric vehicles (HEVs), the energy management strategy (EMS) is essential for their efficiency and energy consumption. Typically, EMS employs rule-based strategies calibrated to general driving conditions. So, this paper proposes to calibrate the EMS of an urban hybrid electric bus that covers a particular route by taking advantage of past driving information. The EMS computes the percentage of the vehicle power demand that must be supplied by each of the sources (fuel and battery) and also controls the heating, ventilating and air conditioning (HVAC) system to achieve cabin thermal comfort. The proposed approach is based on employing an optimal solution by dynamic programing in a previous loop covered by the bus in the considered route. Then, the cost-to-go matrix is stored and used in the following trips by applying the one-step look-ahead rollout, taking profit from the similarities of the loops in the route. To compare and evaluate the performance of the proposed algorithm, a benchmark was carried out by employing the widespread equivalent consumption minimization strategy (ECMS) approach, combined with rule-based strategies in the HVAC control system. Finally, the pareto front presents the trade-off between cabin temperature control performance and total fuel consumption, allowing to compare and evaluate the different EMS calibrations.

The growing call for pollution-free environments has prompted the creation of zero-emission zones (ZEZs) around the world. For regional and national transport, plug-in hybrid electric vehicle (PHEV) are an attractive option, which also offer ZE driving. To address the PHEV challenges of sufficient ZE driving range and of meeting real-world emission targets outside the ZEZs, this work proposes an adaptive supervisory control strategy, which minimizes the total operational costs while complying with tailpipe [Formula: see text] emissions constraints. It combines a Modular Energy Management Strategy (MEMS), for cost-optimal power-split, with an Integrated Emission Management (IEM) strategy for determining the cost-optimal air path setting of the internal combustion engine. A real-time implementable, optimal control strategy is derived based on Pontryagin’s Minimum Principle. To determine the optimal selection of the co-states used in this strategy, a numerical optimization is performed for different route segments and real-world cycles. This study demonstrates that PHEVs can successfully be operated around ZEZs. The best performance is realized with an adaptive supervisory control strategy with different co-states per route segment; compared to the standard strategy with fixed co-states, this proposed strategy was able to achieve cost and [Formula: see text] emission reductions of up to 10% and 22%, respectively, for the studied real-world cycles.

The growing interest in low-impact mobility technologies has elevated the significance of fuel cell electric vehicles (FCEVs) in the automotive sector. Given the complexity of the resulting powertrain, the need for an effective energy management strategy (EMS) becomes essential to optimize efficiency and energy consumption in vehicles with diverse energy sources. Model-based control is the main approach to address the EMS in electrified vehicles. In particular, fuel cell power is commonly represented through a 1D look-up table using the current demand as input to simplify the implementation in a vehicle control unit. Uncertainties that may be implemented in maps due to simplifying hypotheses, dynamics, ageing, etc., can be propagated to powertrain control, motivating the adoption of adaptive look-up tables for FC modelling. In this study, an extended Kalman filter (EKF) is proposed to adapt the look-up table to actual FC behaviour by measuring its power and gradually correcting calibration errors, drift, and ageing. Subsequently, a standard equivalent consumption minimization strategy (ECMS) is employed to control the FCEV. The fuel cell model is calibrated with experimental data from an FCEV. The results demonstrate that the adaptive strategy outperforms the base calibration. Following an extensive simulation campaign, an improvement of 1.1% in fuel consumption was observed. Remarkably, after just one hour of operation, there was a notable 85% reduction in fuel cell power estimation error, even when the EMS was initially fed a biased look-up table.