The goal of this research is to improve the energy productivity of heavy-duty electric vehicles through the development of a predictive cruise controller (PCC), which is optimized for various payload conditions (operation-aware). Energy productivity is defined as the amount of freight transported over a given distance per unit of average energy consumed within a specified travel time. In the trucking industry, both energy consumption and travel time are critical performance metrics. While reducing energy consumption is essential for extending driving range, it must not lead to excessive increases in travel time, which can cause delayed deliveries and higher operational costs. Accordingly, the proposed PCC is designed to be practically feasible for real-world truck operations. The predictive cruise controller is formulated as an optimal control problem (OCP) that simultaneously minimizes energy consumption and reference speed tracking error, with their relative importance governed by a penalty factor. This study demonstrates that the optimal selection of this penalty factor must depend on truck operating conditions, particularly payload variation and travel time constraints, in order to achieve improved energy productivity. The formulated OCP is shown to be convex, which makes it well suited for solution using gradient-based optimization methods. Sequential Quadratic Programming (SQP) is employed to solve the OCP in both simulation and hardware-in-the-loop environments. The results demonstrate that the proposed PCC, combined with an optimal and operation-aware selection of penalty factors, achieves an energy productivity improvement of up to 3.1% while effectively balancing energy consumption and travel time increments. This balanced performance is essential for enabling efficient, reliable, and cost-effective heavy-duty truck operations.

Hybridizing off-highway vehicles remains challenging due to their variable power demands, diverse duty cycles, stringent torque requirements, and limited charging infrastructure. This study investigates the energy management strategy (EMS) for a series-hybrid agricultural tractor. A benchmark charge-sustaining EMS is first developed using dynamic programming (DP) as an offline optimal control technique. These benchmark results are then leveraged to derive rules for a real-time implementable controller, addressing the limitations of DP, such as excessive computation time under highly variable conditions and strong dependence on detailed system models. The proposed charge-sustaining approach is designed to handle multiple operating scenarios and generalize across unseen operations. Validation is performed using experimental duty cycle data from actual field operations. Results show that the proposed approach successfully captures optimal control trends consistent with DP benchmarks, achieving between $\mathbf{3 . 5 \%}$ to $\mathbf{1 8 \%}$ equivalent fuel savings compared to a conventional powertrain for different operations of an agricultural tractor.

HighlightsPowertrain-Aware Explainable Reinforcement Learning Energy Management in Off-Road Vehicles Electrifying high-power off-road vehicles is challenging.Series-hybrid with adaptive EMS addresses operating variability for maximum efficiencyExplainable and interpretable DDQN trained on real off-highway duty cyclesRobust control under uncertainty, unseen, and general operating conditionsUp to 23% fuel savings vs conventional and 11.3% vs rule-based

As Hybrid Electric Vehicles (HEVs) gain traction in heavy-duty trucks, adaptive energy management is critical for minimizing fuel consumption and maintaining battery charge over long operations. To address this, we present a reinforcement learning (RL) framework for engine control in Series HEVs, based on Soft Actor-Critic (SAC) framed as a sequential decision-making problem. Specifically, standard Feedforward networks (FFNs) in the actor and critic are replaced with temporal models, Gated Recurrent Units (GRUs), and Decision Transformers (DTs), to better capture sequential dependencies. Through comprehensive ablation studies, we identify two high-performing and well-generalized sequence-aware agent configurations: a DT actor with a GRU critic (DT-GRU), and a GRU actor with a GRU critic (GRU-GRU). Validation compared these SAC agents against Dynamic Programming (DP) across two training setups: one trained on a single standard cycle (HFET), and another using a wider range of initial SOCs, drive durations, and power demands. On the training cycle (HFET), DT-GRU agent achieved fuel performance within 1.8% of DP, outperforming the GRU-GRU agent, and the FFN-FFN agent at 3.16% and 3.43%, respectively. The unseen highway test cycles, one with significantly higher power demand (US06), while another with extended operation time (HHDDT), compared to the training cycle, allowed us to assess robustness. Both sequence-aware agents showed better generalization than the FFN-FFN agent across these challenging conditions. Notably, agents trained with broader initial conditions not only maintained strong fuel economy but also satisfied critical final constraints.

Reliable estimation of lithium-ion battery capacity degradation is essential for electrified powertrain design and control. This work develops a physics-informed transformer model (PITM) that integrates physics-based degradation trends with multivariate cycling data to learn battery aging dynamics for accurate capacity degradation trajectory estimation across diverse battery chemistries and operating conditions. A physics-based reduced-order model (PB-ROM) is evaluated on 17 cells from five aging datasets to identify degradation-relevant variables and quantify its limits. While the PB-ROM captures long-term degradation patterns, its accuracy degrades under strongly nonlinear conditions, with relative standard error of prediction (RSEP) ranging from 5% to 23%. Guided by these insights, the PITM is developed using dynamic features and trained with a composite loss function that incorporates both experimental capacity measurements and PB-ROM-derived degradation increments. Five single-group PITM models and a generalized multigroup variant (G-PITM) are trained across heterogeneous datasets. The proposed models consistently outperform the PB-ROM, reducing RSEP to as low as 2.96% on unseen test cells and accurately capturing complex degradation behavior. Finally, the PITM is integrated into a series-hybrid range-extender powertrain architecture and evaluated under heavy-duty driving cycles. The PITM demonstrates physically consistent trajectory estimation under realistic load dynamics, confirming its suitability for control-oriented applications.

In this work, the authors introduce and evaluate a series-hybrid range-extender powertrain architecture for high-power off-highway vehicles, addressing key challenges in electrification. This architecture offers precise control of engine operating points but introduces multiple energy conversions that can lead to significant power losses. An effective energy management strategy (EMS) is essential to overcome the losses while minimizing fuel consumption, maintaining battery state of charge (SOC), and ensuring high engine efficiency. However, the highly variable, stochastic, and uncertain operating conditions typical of off-highway applications limit the effectiveness of threshold-rule-based control strategies and reduce the optimality of model-based optimal control strategies. To address these challenges, this work proposes an explainable, interpretable reinforcement-learning-based EMS that adapts to high-variance environments, handles uncertainty, and generalizes to unseen operating conditions. A powertrain constraints-aware, explainable, and interpretable Double Deep Q-Network (XI-DDQN) controller is devel-oped, trained on actual experimental duty cycles, and evaluated on unseen actual real-world scenarios to assess its adaptability and generalization. The proposed EMS is benchmarked against dynamic programming (DP) and compared to a rule-based controller. The results, based on the validated model, demonstrate that the proposed XI-DDQN successfully captures the optimal control behavior of DP while maintaining robust adaptability across uncertain and generalized operating conditions. The XI-DDQN-based EMS achieves up to a 23% reduction in fuel consumption compared with a conventional powertrain and delivers an average 11.7% reduction relative to the rule-based strategy under unseen operating conditions. These findings highlight the effectiveness of the proposed approach in achieving fuel-efficient, robust, and interpretable energy management for next-generation hybrid off-highway machinery.

This article introduces a comprehensive cooperative navigation algorithm to improve vehicular system safety and efficiency. The algorithm employs surrogate optimization to prevent collisions with cooperative cruise control and lane-keeping functionalities. These strategies address real-world traffic challenges. The dynamic model supports precise prediction and optimization within the MPC framework, enabling effective real-time decision-making for collision avoidance. The critical component of the algorithm incorporates multiple parameters such as relative vehicle positions, velocities, and safety margins to ensure optimal and safe navigation. In the cybersecurity evaluation, the four scenarios explore the system’s response to different types of cyberattacks, including data manipulation, signal interference, and spoofing. These scenarios test the algorithm’s ability to detect and mitigate the effects of malicious disruptions. Evaluate how well the system can maintain stability and avoid collisions under compromised conditions. It also analyzes the impact of varying levels of attack severity on overall system performance. The cooperative navigation framework highlights its potential as a robust solution for secure, efficient, and safe autonomous vehicle operations in increasingly interconnected and potentially hostile environments. Case 1 simulates communication jamming, where all channels except vehicle-to-vehicle communication are compromised. Case 2 extends this to jamming in the smart traffic light system, creating a non-signalized environment. Case 3 represents an ideal scenario with seamless communication. Case 4 explores vulnerability to deliberate interference in actor vehicle velocities, amplifying collision risk. Surrogate optimization with radial functions ensures proactive collision avoidance, while model predictive control with the interior point solver optimizes trajectory planning, promoting collision-free operation, and improving traffic flow. The algorithm’s outputs are seamlessly integrated into the vehicle control system, with the ego vehicle’s dynamics modeled realistically. Through extensive simulations, the algorithm proves effective across diverse scenarios, including communication disruptions and intentional interference. The research contributes to cooperative navigation system advancement, showcasing potential improvements in safety, efficiency, and adaptability in contemporary vehicular environments. The algorithm’s ability to handle various scenarios presents promising prospects for future intelligent transportation systems research.

We address range anxiety in multi e-axle heavy-duty electric vehicles through optimal torque management. A primary cause of reduced driving range and increased energy consumption in heavy-duty electric vehicles is the frequent demand for peak driving torques from the e-axles due to significant payload variations, which can increase the vehicle weight by up to four times its curb weight. An additional challenge lies in the delivery of the requested driving torque from multiple e-axles. Conventional powertrain controllers for multi e-axles evenly distribute the requested driving torque, which often proves suboptimal. It is critical to distribute requested driving torque optimally for low power requests when the e-axles operate in low-efficiency regions. This research work presents a two-step optimal torque management strategy to increase the range. Step 1 utilizes look-ahead (road grade and speed limit) and truck operation (payload) information to generate an optimal driving torque profile for e-axles, particularly suited for highway driving conditions. Step 2 optimally distributes torque across all e-axles to minimize powertrain losses while ensuring smooth operational transitions. The formulated optimal control problems for step 1 and step 2 are solved using Sequential Quadratic Programming (SQP) in both simulation and hardware-in-the-loop (HiL) environments to quantify the performance improvements relative to the baseline, defined as a conventional cruise controller with even torque distribution. The HiL setup employs a Speedgoat Baseline Real-Time Target as the optimal controller and a dSPACE Scalexio unit as the validated digital twin of an actual truck. Dynamic Programming (DP) and a brute-force method establish benchmark performance. Both simulation and HiL results are comparable and show significant gains in both range and freight efficiency. The proposed strategy yields a range improvement of 16.9%+/- 5.0% under no-payload conditions and 4.9%+/- 3.1% under full-payload conditions, while enhancing freight efficiency by 10.5%+/- 1.8%. The results demonstrate significant improvements for heavy-duty electric truck operations.

This study investigates depot charging capacity planning and charging schedule optimization for heavy-duty electric vehicle fleets, explicitly incorporating grid-aware infrastructure considerations. As fleet electrification expands, uncoordinated charging demand can intensify peak loads, increase electricity costs, and impose additional stress on power systems, underscoring the need for integrated planning approaches. A two-phase optimization framework is proposed that jointly determines depot charging infrastructure capacity and time-dependent charging schedules by accounting for peak demand charges, infrastructure installation costs, and operational expenses, while incorporating temporal fleet usage patterns and grid load conditions. A case study is conducted to evaluate optimal charger capacity portfolios and operational schedules under representative tariff structures. The results demonstrate that coordinated depot charging can substantially reduce peak demand and total system costs, including both infrastructure investment and operating expenses, compared with uncoordinated charging strategies. While scenarios with higher power charging infrastructure lead to increased peak demand charges, these impacts can be effectively mitigated through optimized scheduling. Overall, the findings highlight that grid-aware depot charging planning enhances both cost efficiency and power system resilience, providing quantitative evidence to support infrastructure investment decisions and policy design for large-scale fleet electrification.

This study develops an aging-aware deep reinforcement learning framework for heavy-duty electrified powertrains, optimizing fuel consumption jointly with three physics-based degradation dynamics: battery calendar and cycling aging and generator thermal aging. The discrete–continuous actions are addressed using Physics-Aware Soft Actor–Critic (PA-SAC) agents, and training stability is enhanced by a Dynamic Programming (DP)-based warm start combining behavior cloning, expert replay-buffer seeding, and critic warmup. The framework is applied to a range-extender series-hybrid architecture, with agents trained on the Heavy Heavy-Duty Diesel Truck (HHDDT) cycle and evaluated on the training cycle and two unseen test cycles under a terminal state-of-charge (SOC) constraint of [0.5480, 0.5520]. Benchmarked across five seeds per case against DP optima, Twin Delayed Deep Deterministic Policy Gradient (TD3), and Proximal Policy Optimization with Lagrangian constraints (PPO-L), PA-SAC attains the tightest DP gap on the training cycle and lowest seed-to-seed variability. Aging objectives reshape the control strategy, reducing battery capacity degradation due to cycling aging by up to 91% relative to the fuel-only baseline and redistributing load between the battery and engine-generator subsystems under the aging-aware objective. The trained policy is deployed on an embedded controller target in a controller hardware-in-the-loop testbench, reproducing the training-time control behavior across twelve runs.

This study explores depot charging scheduling for heavy-duty electric fleets, emphasizing energy dynamics and grid resilience. From a demand-side management (DSM) perspective, it integrates grid-aware planning by analyzing the energy flow from the grid to charging infrastructure and its impact on recharging strategies. To optimize charging times while enhancing grid stability, the study proposes a recharging scheduling framework that incorporates key metrics such as peak shaving and alignment with grid interchange data. The methodology is built on a metaheuristic Tabu Search algorithm, designed to improve grid resilience and mitigate peak demand impacts. A case study demonstrates the effectiveness of the proposed framework compared to conventional approaches that do not consider grid stability. These findings offer valuable insights for policymakers, grid operators, and fleet managers, aiding in the development of resilient and efficient large-scale electrified fleet operations.

Agricultural tractors contribute significantly to emissions, yet their electrification faces challenges due to their versatile operations and varying terrain conditions. This study explores a series hybrid range extender powertrain architecture to benefit from controlling the ICE operating points to optimize and minimize fuel consumption. A high-fidelity framework is developed that allows for exploring the benefits of the proposed powertrain architecture. A charge-sustaining energy management strategy (EMS) is implemented to optimize the power split between two power sources: (1) a Power Generation Unit (PGU) and (2) a battery pack. This strategy also allows the ICE to operate at the optimal operating points and prevent battery depletion throughout the tractor operation while minimizing fuel consumption. The initial results decrease the average fuel consumption rate by 50% compared to the baseline conventional powertrain for agricultural tractors, in addition to the ability to downsize the engine from 500 kW to 275 kW. 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/)

Long-haul truck drivers are mandated to take off-duty time of 10 h (a.k.a. hoteling) before driving. During the hotel phase, drivers spend time inside their trucks (sleeper cabs) and idle the internal combustion engine for comfort by utilizing the heating, ventilation, air-conditioning (HVAC), and other onboard appliances. For one 10-h period, the average cost is about $40, which can be a lot when considering a million truck drivers idling overnight. SuperTruck II is a 48 V mild-hybrid heavyduty truck with auxiliary loads powered by an onboard battery pack. An optimal control algorithm is developed to charge the battery pack during the drive phase up to a certain state-of-charge (SOC) level, sufficient to meet the power demands of the auxiliary load during the hotel phase. This article captures the research done to predict energy consumption in a mild-hybrid heavy-duty sleeper truck during hoteling. Physics-based gray box models are developed to estimate the power consumption of an electronically controlled compressor. For other auxiliary loads, a machine learning algorithm is developed to predict the power as a time series by tracking the user activity. The developed physics and data-driven models are validated with experimental data from heavy-duty trucks to show their efficacy. These validated models generate precise load profiles fed to the developed dynamic programming framework to generate the optimal SOC trajectories. These models help the vehicle battery pack charge only up to the SOC necessary for the hotel phase during the drive time. When the vehicle is out of charge during the hotel phase, these models also help in estimating the amount of idling required to charge the battery enough to support the rest of the hotel period. This saves unnecessary idling. As a result, a cost savings of $40 and CO2 reduction of 175 lb to the environment is achieved for a single heavy-duty truck with a sleeper cab during the hoteling phase.

Effective traffic management and energy-saving techniques are increasingly needed as metropolitan areas grow and traffic volumes rise. This work estimates fuel consumption over three selected routes in an urban context using spatio-temporal modeling essentially building on a previously developed approach in traffic prediction and forecasting. A weighted adjacency matrix for a Graph Neural Network (GNN) is constructed in the original approach which combines graph theory frameworks with travel times obtained from average speeds and distances between traffic count stations. Next, the traffic flow estimate uncertainty is measured using Adaptive Conformal Prediction (ACP) to provide a more reliable forecast. This work predicts fuel consumption under different scenarios by utilizing Monte Carlo simulations based on the expected traffic flows providing insights into energy efficiency and the best routes to take. The study compares passenger vehicles' and heavy-duty trucks' mean fuel consumption under morning and evening traffic conditions. For passenger vehicles, the predicted fuel consumption showed a maximum error of 5.6% when compared to observed values, while for heavy-duty trucks, the maximum error was 8.4%. The model's capacity to effectively represent temporal fluctuations in traffic patterns and their effects on fuel economy is demonstrated by this comparative analysis. The study shows the practical applicability of this approach for energy-efficient route planning and urban traffic management by validating the estimated fuel consumption against real-world data. This gives transportation planners a comprehensive tool to help them make decisions that minimize environmental impacts and maximize fuel efficiency.

Heavy-duty trucks idling during the hotel period consume millions of gallons of diesel/fuel a year, negatively impacting the economy and environment. To avoid engine idling during the hotel period, the heating, ventilation, and air-conditioning (HVAC) and auxiliary loads are supplied by a 48 V onboard battery pack. The onboard battery pack is charged during the drive phase of a composite drive cycle, which comprises both drive and hotel phases, using the transmission-mounted electric machine (EM) and battery system. This is accomplished by recapturing energy from the wheels and supplementing it with energy from the engine when wheel energy alone is insufficient to achieve the desired battery state of charge (SOC). This onboard battery pack is charged using the transmission-mounted EM and battery system during the drive phase of a composite drive cycle (i.e., drive phase and hotel phase). This is achieved by recapturing wheel energy and energy from the engine when the wheel energy is insufficient to achieve the desired SOC during the drive phase. In the authors’ previous work, a dynamic programming (DP)–based framework is developed that employs a multi-objective cost function to minimize fuel consumption and maximize the regeneration to achieve the benchmark results for the SOC trajectories. This article discusses the real-time implementable control strategies for the heavy-duty truck’s hybrid powertrain, including the mode switch and EM torque for charging. The mode switch is a rule-based control strategy that responds to the wheel torque demand, while the EM torque’s control can have several approaches, such as rule-based, optimal charging strategies that are inspired by equivalent cost minimization strategy (ECMS), or adaptive strategy that updates the equivalent factor according to the battery SOC state. This work presents and studies the different choices to control the EM torque and their impact on vehicle performance and energy consumption. The complete cycle results are compared with the benchmark results, and the energy analysis is accomplished to validate the efficacy of the proposed real-time implementable optimal control strategies (i.e., rule-based and adaptive ECMS). The adaptive optimal control strategy is the potential candidate to be implemented on a real heavy-duty vehicle for optimal management of hotel loads, as it produces the SOC trajectory closer to the benchmark results within the error of ±1.25% while costing minimal fuel consumption. The fuel saving of 2.96% is achieved when compared to conventional heavy-duty trucks for each day of a typical highway trip and hotel phase for each heavy-duty truck, which is 18.2% higher than the rule-based control strategy.

Heavy-duty vehicles (i.e. class-8 trucks) are the backbone of the freight movement in the United States. The electrification of heavy-duty vehicles brings huge benefits in the form of reduced greenhouse Gas (GHG) emissions, reduced annual energy consumption, and annual fuel costs. The efficiency improvement of battery-powered electric trucks will make them more attractive as compared to conventional counterparts. Battery-powered electric trucks have dual e-axle-based powertrains to fulfill the traction power requirements. Currently, the dual e-axle control is based on the even torque split which fails to provide high efficiency at low power requests. This research paper proposes the optimal torque allocation technique for e-axle to minimize traction losses. It leads to a significant improvement in the overall energy efficiency of the powertrain. The optimal torque allocation routine is based on sequential quadratic programming(SQP) with constraints on torque split ratio range, gear number on axles, and battery State of Charge (SOC). The proposed technique is validated and compared with the conventional/even torque split strategy. Results show that energy losses at e-axle are reduced by 6.15% and overall energy efficiency is increased by 0.682%.

This study evaluates the robustness and sensitivity of a Reduced-order battery aging model and a Dynamic thermal aging model for electric machines (EMs) in heavy-duty electric vehicles. The battery aging model is calibrated and validated using two datasets, with consideration of two temperature conditions and six charging/discharging rates (C/4, 1C, 3C at 23 degrees C, and 1C, 4C, 8C at 30 degrees C). The model estimates capacity degradation with good accuracy under both fast and normal charging scenarios, effectively capturing the degradation trends observed in experimental data. The estimated capacity degradation resulted in a best-case Relative Standard Error of Prediction (RSEP) of 6.21% and a worst-case RSEP of 14.25%. For the EM, four temperature profiles are utilized: two obtained experimentally by testing an EM under different conditions, and two generated by inputting distinct drive cycles into a range extender powertrain simulator. Using the experimental data, the Dynamic thermal aging model indicates a significant reduction in EM lifetime when exposed to prolonged high temperatures. For the simulated temperature profiles, the Federal Highway Driving Schedule (FHDS) results in a notable increase in lifetime loss due to extended exposure to elevated temperatures, whereas the High-Efficiency Truck Users Forum (HTUF) drive cycle shows a lower lifetime loss, attributed to reduced thermal stress.

This study addresses the challenges of electrifying heavy-duty vehicle fleets, particularly school buses, by focusing on the development of dedicated depot charging infrastructure and grid resilience. A key challenge is managing recharging limitations while considering grid resilience in the electrification of school bus fleets. Using real operational data, the study introduces a two-phase approach to optimize both charging infrastructure and scheduling. In the first phase, the optimal number of chargers is determined to ensure sustainable fleet operations. In the second phase, charging schedules are refined to reduce peak power demand and improve grid resilience. Experimental results demonstrate that approximately half the fleet size is required in chargers, with distributed charging and peak shaving strategies reducing peak power demand by 20% to nearly 45%. These findings offer practical insights for fleet managers, grid operators, and policymakers on enhancing grid resilience and managing electrified fleets effectively.

Off-highway vehicles, with their unique requirements of durability, high power, and torque density, are typically powered by diesel ignition internal combustion engines (ICEs). This reliance on ICEs significantly contributes to greenhouse gases (GHGs) emissions. For this reason, there is an urge to develop an energy-efficient powertrain architecture that produces fewer GHGs emissions while meeting the variable torque levels and variable speeds and performing various duty cycles with high efficiency. In order to select the energy-efficient powertrain architecture for the off-highway vehicle, different existing powertrain architectures (i.e., series hybrid, parallel hybrid, series-parallel hybrid, conventional) for off-highway applications have been studied to highlight their pros and cons. This is done considering the different duty cycles and applications along with Life Cycle Analysis (LCA). Off-highway vehicles operate under different road/surface conditions than on-road vehicles, which affects the powertrain’s performance. Hence, the terrain properties are also discussed and considered in this work to select the appropriate powertrain architecture. The selected powertrain architecture takes into account the above-mentioned loads to ensure better performance. Lastly, the authors present the details of the proposed powertrain architecture for off-highway vehicles in this manuscript, including its components, architecture, and modes of operation. The proposed architecture is compared with the conventional off-road vehicle’s architecture to highlight its benefits, including energy benefits and fuel consumption benefits, that lead to GHG emissions reduction benefits.