Provides society information that may include news, reviews or technical notes that should be of interest to practitioners and researchers.
Transportation electrification is transforming power systems, introducing both new stresses and unprecedented flexibility. While large-scale electric vehicle (EV) adoption can amplify peak demand, the aggregated storage capacity of EV batteries offers a powerful opportunity to support grid stability through vehicle-to-grid (V2G) technology. Realizing this vision, however, requires coordinating millions of mobile, user-constrained assets in real time, an inherently complex challenge that exceeds the capabilities of traditional centralized control. This article presents a forward-looking deployment vision for a grid that drives itself through an AI-orchestrated framework based on a cyber-physical-privacy triad. By integrating multi-agent reinforcement learning (MARL) for autonomous coordination, federated learning for privacy-preserving data protection, and control barrier functions (CBFs) for hardware-level safety, we present a path toward a truly resilient virtual power plant. We map these intelligence layers directly to industry standards, such as the ISO 15118-20 Dynamic Mode handshake, to bridge the gap between high-level market optimization and sub-second power electronics control. From managing electrochemical battery health to mitigating coordinated load-altering cyber-attacks, the proposed architecture offers an engineering blueprint for a self-healing energy ecosystem. Ultimately, this convergence allows the global EV fleet to operate as a cohesive energy resource, accelerating the transition to resilient, decarbonized energy infrastructure.
This article presents an outlook for onboard power conversion as the core of modern marine power and propulsion systems. Initially, it highlights the motivation behind the transition toward more efficient and sustainable shipboard technologies and introduces key electrification and hybridization solutions. To support this progression, a classification of power converters is presented along with an overview of emerging technologies, including solid-state transformers (SSTs). Building on this foundation, a descriptive review of shipboard power architectures is provided, supported by single-line diagrams, highlighting the current trend to shift from conventional AC architectures toward low-voltage DC (LVDC) and medium-voltage DC (MVDC) topologies, where DC grids act as key enablers of flexible onboard systems. Furthermore, insights are provided into the system integration challenges of alternative marine energy sources and storage technologies, paving the way for long-term decarbonization in the maritime sector.
As the maritime sector undertakes a US${\$}$3.7 trillion transition toward net-zero greenhouse gas emissions, electrification is limited no longer by individual technologies but by the ability to coordinate decisions across tightly coupled ship–port systems. This article examines maritime electrification through the lens of Maritime 5.0, where artificial intelligence (AI) is positioned as a bounded, human-centered planning and decision support layer operating above certified control and protection systems. We first identify the key coordination challenges that emerge as vessels and ports electrify at scale, including design uncertainty over long asset lifetimes, high-power ship–port integration, operational complexity across modes, and transparency frictions in compliance and economics. Building on this diagnosis, we introduce a set of AI-enabled capabilities that support uncertainty-aware forecasting, rapid scenario screening, structured design exploration, human-in-the-loop operations, and traceable decision justification. A corridor-scale case study illustrates how these capabilities can be integrated incrementally into real maritime workflows. Together, these insights show how AI-empowered decision support can reduce coordination friction in electrified ship–port systems while preserving safety, accountability, and the human-centered principles central to Maritime 5.0.
Vehicle-to-grid (V2G) technology is rapidly transforming electric vehicles (EVs) from passive loads into distributed, dispatchable grid assets. As renewable penetration increases and distribution networks evolve toward bidirectional power flow, coordinated control of EV fleets becomes essential for maintaining stability, reliability, and economic efficiency. However, scalable V2G deployment presents complex challenges across converter control, battery lifecycle management, cybersecurity, interoperability, and market participation. This article presents a system-level framework for artificial intelligence (AI)-enabled V2G integration at the grid edge. Layered architecture is proposed that integrates forecasting engines, degradation-aware optimization, fleet aggregation strategies, and grid-forming/grid-following control schemes. The framework incorporates transformer thermal constraints, voltage regulation limits, and dynamic market signals while preserving battery state of health and warranty compliance. The role of edge intelligence versus centralized cloud coordination is examined, highlighting tradeoffs in latency, resilience, and scalability. Interoperability considerations across communication standards and cybersecurity resilience are analyzed, with emphasis on secure bidirectional energy transactions and coordinated fleet dispatch. The article further explores how V2G-enabled EV fleets can function as virtual power plants, supporting frequency regulation, peak shaving, renewable smoothing, and microgrid islanding during extreme events. Drawing from emerging pilot deployments and industry lessons learned, this work outlines the technical and regulatory pathways necessary to transition V2G from demonstration to large-scale infrastructure. By integrating advanced control theory, AI, and grid modernization strategies, V2G systems can become a foundational component of resilient, low-carbon power networks.
The electrification of Class 7–8 trucks is advancing from pilot deployments toward broader commercial operation. The primary constraint has shifted from vehicle technology to charging infrastructure and grid integration. This article examines the engineering challenges of scaling heavy-duty battery-electric truck charging, with emphasis on depot-centric megawatt-scale loads, high-power public charging for long-haul applications, and the evolving relationship between charger capacity and fleet utilization. Drawing on recent scenario-based infrastructure modeling and early deployment experience, the discussion highlights key considerations related to charger-to-truck ratios, infrastructure cost drivers, utility interconnection timelines, and operational reliability. The article further identifies opportunities for standards development, interoperability, power quality management, and technical education as megawatt-class charging systems enter commercial service. These insights are intended to inform engineers, utilities, fleet operators, and infrastructure developers working to enable scalable, reliable charging solutions for heavy-duty battery-electric trucks.
Vehicle-to-grid (V2G) is moving from isolated demonstrations toward early commercial deployment, but the control architecture of the charger will strongly shape where V2G can be deployed safely and where it can create real added value. Most operational V2G systems today are grid-following (GFL), which works well in stiff grids and dispatchable export services. However, several high-value V2G applications, including weak-feeder voltage support, islandable microgrids, and off-grid or energy-access settings, arise where the grid reference is weak, distorted, or absent, linking V2G deployment to wider goals for affordable clean energy, resilient infrastructure, and sustainable communities. In these conditions, grid-forming (GFM) control becomes attractive because the charger can help establish local voltage and frequency rather than simply follow them. This article provides a deployment-oriented comparison of GFL and GFM V2G, with emphasis on offboard bidirectional chargers, aggregated fleets, multisocket charging facilities, and the distinction between single-phase and three-phase deployment contexts. It also clarifies the maturity of current technologies, noting that GFL V2G is closer to early commercial use, while fully integrated dual-mode GFL/GFM V2G remains at the development stage. The core message of this article is that V2G should not be designed around a single control mode. GFL remains the practical choice for most near-term grid-connected services, while GFM becomes valuable where the charger must support a weak, islanded, or absent grid reference. The future pathway is therefore dual-mode, service-aware V2G, where chargers and fleets select the appropriate control behavior according to grid condition, site architecture, user constraints, and service requirements.
A usable vehicle-to-grid (V2G) system is not defined by a single converter or a single protocol. It takes shape only when charger hardware, communication between the electric vehicle (EV) and the electric vehicle supply equipment (EV-EVSE), charger management, and utility-side coordination work together coherently in the field. V2G devices are moving beyond isolated pilots toward a more standards-based infrastructure stack, yet real-world deployment still hinges on coordination across charger hardware, vehicle–charger communication, middleware, grid interfaces, and utility operating requirements. This article outlines the system-level configuration of that stack and traces the connections among bidirectional charger capability, ISO 15118-20, applications on charger-management platforms, and IEC 61850-based utility integration. We do not imply that all layers have matured at the same pace. Instead, the aim is to show where public evidence is already available, where standards have progressed faster than deployment, and why fleet and depot applications continue to lead real-world demonstrations. Results from school bus and fleet programs suggest that the architecture can already provide useful services in controlled operating contexts. Even so, the overall picture remains mixed. The technical basis for bidirectional charging is now strong, but routine multivendor deployment still depends on better interoperability, stronger utility-facing integration, and greater confidence in commercial operating models.
DC microgrids are increasingly adopted in shipboard and industrial power systems due to their efficiency and compatibility with power-electronic sources, storage, and loads. Conventional power management typically treats DC-link voltage deviation as an error to be eliminated through tight regulation or supervisory coordination. This article presents an alternative variable voltage-based power management paradigm in which controlled DC-link voltage variation—arising primarily from droop-controlled voltage sources—is embraced as an inherent and informative consequence of system loading and network behavior. Loads operate conventionally over a wide nominal voltage range and remain uninvolved in power management unless predefined under- or overvoltage thresholds are exceeded. Upon threshold activation, load-side converters transition from current/power-regulated operation to voltage-regulating operation, with the resulting current determined solely by external system conditions rather than by explicit current commands. The same interface-level principles govern both normal operation and boundary conditions, including source saturation, source loss, regenerative power injection, and large load rejection, without explicit event detection, centralized load shedding, or communication-dependent coordination. In shipboard DC systems, this yields deterministic and certifiable power prioritization, reduced reliance on communication infrastructure, symmetric handling of undervoltage and overvoltage conditions, and inherent contribution to blackout prevention.
The transition to electric mobility offers an unprecedented opportunity to leverage electric vehicles (EVs) as distributed energy resources through vehicle-to-grid (V2G) technology. However, the commercial barrier to V2G is no longer just interoperability; it is the lack of trust regarding battery degradation and fair compensation. Current V2G dispatch strategies rely predominantly on state of charge (SoC) as the sole control variable, treating the battery as a fixed-power product rather than a state-dependent flexibility product. This article argues that SoC-only control is fundamentally insufficient and proposes a dynamic, battery-aware operating envelope that defines safe power limits for both discharging (V2G) and charging (G2V) based on a real-time impedance, equivalent circuit model (ECM) parameters, temperature, SoC window, and state of health (SoH). By integrating online impedance measurement and robust estimation algorithms, such as an extended Kalman filter (EKF)-based state estimation directly into the bidirectional charger, the battery becomes fully observable at the point of service. This framework shifts the narrative from “V2G degrades batteries” to “V2G should be dispatched only inside a safe flexibility envelope, ” enabling warranty-conscious compensation, consumer trust, and the evolution toward cloud-assisted digital-twin fleet control.
The growing complexity of grids challenges reliability. And the complexity is increasing in grids ranging from the legacy grid, to microgrids, to the power grids used in electrified transportation. It is not surprising that achieving here-to-fore unneeded levels of reliability can increase costs. A grid, however, can tell us if it is at risk. But innovative approaches to measurements can take advantage of the available information. The legacy approach to situational awareness in the grid has been to make narrow bandwidth measurements in substations. Relays have sensors to conduct these measurements to permit rapid disconnection of power when a significant problem arises. Emerging technology, however, is augmenting this legacy approach with wide bandwidth distributed sensors. These devices capture information from the power system that was previously unnecessary when reliability requirements were lower. The new systems combine novel sensors with edge computing and machine learning to exploit the available information to achieve reliability.
Electric bus fleets are expanding worldwide as cities seek to reduce emissions and improve air quality, reshaping urban systems. Unlike private electric vehicles, transit fleets operate on predictable schedules, have centralized charging infrastructure, and feature large battery capacities, making them uniquely suited to support the power grid. Through vehicle-to-grid (V2G) technology, these fleets can serve not only as transportation assets but also as distributed energy resources (DERs), providing services such as peak shaving, renewable energy balancing, and ancillary grid support. However, realizing this potential requires coordinated charging strategies, integration with electricity markets, and careful management of battery degradation. This article presents a system-level perspective on the role of electric bus fleets in future grid-interactive transportation systems. It introduces the concept of electric bus aggregators as coordinating entities between transit operators and power systems and discusses smart charging strategies that enable fleets to participate in energy markets while maintaining reliable transportation operations. By bridging transportation electrification and power system flexibility, electric bus fleets may play a critical role in developing more resilient and sustainable urban energy infrastructures.
Electric vehicles (EVs) are very popular these days. Their adoption transforms transportation systems and brings in new operational challenges for modern power grids. Vehicle-to-Grid (V2G) ecosystems enable EVs to function as distributed energy resources supporting grid stability and flexibility. However, current V2G implementations face coordination inefficiencies, cybersecurity vulnerabilities, and fragmented market participation frameworks. These limitations restrict large-scale integration of bidirectional charging infrastructure across heterogeneous grid environments. Artificial intelligence (AI) offers advanced capabilities for forecasting, adaptive control, and real-time decision-making in complex energy networks. AI-driven control systems can coordinate thousands of EVs while balancing grid demand, renewable variability, and user mobility requirements. AI also enables predictive battery management, minimizing degradation while maximizing economic participation in energy markets. Cyber-resilient architectures become essential as V2G networks expand digital connectivity between vehicles, aggregators, utilities, and market platforms. Robust security frameworks protect communication channels, safeguard user data, and maintain operational reliability under evolving cyber threats. Market integration mechanisms allow aggregated EV fleets to participate in ancillary services, demand response, and energy trading. This study examines AI-enabled orchestration of V2G ecosystems combining intelligent control strategies, cyber-resilient architectures, and market-aware optimization.
Global transportation electrification is accelerating rapidly, turning heavy-duty electric vehicles from passive grid loads into massive, distributed energy resources. However, expanding traditional, hardwired grid infrastructure to connect geographically isolated or congested microgrids remains cost-prohibitive and environmentally challenging. Furthermore, conventional vehicle-to-grid (V2G) approaches often suffer from unpredictable vehicle availability and severe battery degradation. This article presents a system-level paradigm shift driven by an AI-enabled energy management system (EMS): orchestrating diverse transit fleets, such as school, corporate shuttle, and public buses, as a synchronized, mobile energy infrastructure. By leveraging this AI orchestrator and the highly predictable, staggered dwell times of these fleets, virtual power lines can be established to physically transport energy from areas of renewable surplus to regions of deficit. The discussion explores AI-managed practical innovations, including health-aware shallow cycling to equalize battery degradation across the fleet, and the life-saving potential of grid-forming inverters that transform parked buses into rescue microgrids during catastrophic blackouts. Ultimately, by monetizing these automated energy services, municipalities can unlock new economic models to subsidize public transit fares, proving that the future of grid resilience is not just stationary, it is mobile.
Unsupervised domain adaptation (UDA) is critical for generalizing traffic scene understanding across diverse environments where manual data annotation is not feasible. Modern object detection models frequently fail during deployment due to domain shift, in which detection performance drops due to variables such as weather, lighting, and urban architecture. This paper examines strategies to bridge this gap by encouraging models to learn from unlabeled target data through self-training. In this approach, the model generates pseudo-labels from its own high-confidence predictions to supervise optimization. To ensure stability in this process, we analyze the teacher-student framework, which defines a slow-updating teacher model to provide reliable guidance and mitigate error reinforcement due to noisy pseudo-labels. Furthermore, we explore feature alignment methods, such as adversarial, contrastive, and optimal transport methods, to learn domain-invariant semantics over superficial appearance cues. Through the integration of these methods, a framework is developed to build robust traffic scene understanding systems with high accuracy in diverse environments.
The electrification of public and commercial transportation is driving a rapid evolution in charging technologies capable of delivering high power safely and reliably with minimal human intervention. Pantograph-based conductive charging systems have emerged as a promising option for electric buses, heavy-duty trucks, and fleet vehicles that require fast turnaround times and minimal driver intervention. This article provides a comprehensive review of pantograph charging systems, explaining their working principles, configurations, power electronics, communication protocols, and integration with vehicle fleets and the electric grid. Current global deployments are reviewed, highlighting trends in standardization, interoperability, and safety. Finally, this article explores the future scope, including megawatt-level charging, renewable integration, and digitalized infrastructure for smart cities and commercial transportation.
The recent emergence of large language models (LLMs) such as GPT-3 has marked a significant paradigm shift in machine learning. Trained on massive corpora of data, these models demonstrate remarkable capabilities in language understanding, generation, summarization, and reasoning, transforming how intelligent systems process and interact with human language. Although LLMs may still seem like a recent breakthrough, the field is already witnessing the rise of a new and more general category: multimodal, multitask foundation models (M3T FMs). These models go beyond language and can process heterogeneous data types/modalities, such as time-series measurements, audio, imagery, tabular records, and unstructured logs, while supporting a broad range of downstream tasks spanning forecasting, classification, control, and retrieval. When combined with federated learning (FL), they give rise to M3T Federated Foundation Models (FedFMs): a highly recent and largely unexplored class of models that enable scalable, privacy-preserving model training/fine-tuning across distributed data sources. In this paper, we take one of the first steps toward introducing these models to the power systems research community by offering a bidirectional perspective: (i) M3T FedFMs for smart grids and (ii) smart grids for FedFMs. In the former, we explore how M3T FedFMs can enhance key grid functions, such as load/demand forecasting and fault detection, by learning from distributed, heterogeneous data available at the grid edge in a privacy-preserving manner. In the latter, we investigate how the constraints and structure of smart grids, spanning energy, communication, and regulatory dimensions, shape the design, training, and deployment of M3T FedFMs.
