Wireless Networks Mobile Electric Vehicles Miao Wang Ran Zhang Xuemin (Sherman) Shen Online Charging and Discharging Wireless Networks Series Editor Xuemin Sherman Shen University of Waterloo Waterloo, Ontario, Canada More information about this series at http www springer comseries14180 http www springer comseries 14180 Miao Wang • Ran Zhang • Xuemin (Sherman) Shen Mobile Electric Vehicles Online Charging and Discharging 123 Miao Wang University of Waterloo Waterloo, ON, Canada Xuemin (.
Introduction to the Smart Grid
Global energy demand is soaring, placing a heavy burden on existing resources and the power grid while driving environmental pollution and climate change Renewable energy sources, particularly solar power, photovoltaic systems, and wind energy, have become increasingly attractive as vital supply options to reduce pollution The adoption of non-conventional renewables enables distributed generation (DG) that can be connected locally at the distribution system level With higher DG penetration, improving the efficiency, reliability, and sustainability of electricity production and distribution has spurred the exploration of smart grids, which leverage communication technologies to monitor and respond to the behaviors of both suppliers and consumers.
M Wang et al., Mobile Electric Vehicles, Wireless Networks,
Fig 1.1 Architecture of power grid penetration of renewable energy sources in DGs, the power quality, reliability and security standards, power electronic interfaces and controls need to be studied in the smart grid Specifically, due to the instability of the renewable energy supply, fluctuations on frequency and voltage at buses are introduced into the smart grid.
In the existing power grid, the main components are shown in Fig.1.1 Gen- erations are the resources to generate the energy for the whole systems Through transmission, substation, and distribution parts, energy can be delivered to the customer ends The existing electricity power grid is unidirectional based on one- way communication system Due to the one-way communication system in the existing power grid, the utilities or generations do not take corrective actions based on the information received from the meters Thus, there are many challenging issues in the existing power grid, including generation diversification, demand response, and the problem of carbon footprint, and so on.
The next-generation electricity grid, also called the smart grid or intelligent grid, integrates smart meters, distributed control, and two-way communication with customers to enable parameter adjustments In the smart grid, technologies, concepts, topologies, and approaches for generation, transmission, and distribution can be included, while providing pervasive control and monitoring of grid operations This evolution represents the convergence of information technology and communication technology with power system engineering to enable integrated control and monitoring across the entire grid.
Within smart grids, generation will prioritize cleaner, more efficient bulk-generation technologies to access high-quality renewable energy through transmission, minimize wide-area disturbances, and address congestion in future networks On the distribution side, the system will accommodate new end-user technologies and greater consumer participation—such as plug-in electric vehicles (PEVs), distributed generations (DGs), smart loads, and microgrids—enabling more efficient load management and distributed generation, with PEVs serving as a key component for both energy storage and controllable loads in the smart grid.
An Overview of EVs and Smart Charging in Smart Grid
Plug-in electric vehicles (PEVs) with energy storage and controllable loads can help the electric grid compensate for the fluctuations of renewable generation at distributed generators (DGs) The stored energy in certain PEVs can be discharged to meet load demand or to charge other PEVs, reducing the need for additional storage capacity With bi-directional charging, PEVs can both draw energy from the grid (G2V) and send energy back to the grid (V2G) Through V2G, the discharged energy from PEVs supports grid frequency and voltage regulation Furthermore, an aggregator can coordinate charging and discharging across a group of PEVs, enabling energy transfer among PEVs at swapping stations via vehicle-to-vehicle (V2V) charging/discharging.
1.2 An Overview of EVs and Smart Charging in Smart Grid
Plug-in electric vehicles (PEVs) have become a promising component of sustainable and eco-friendly transportation systems, attracting increasing attention Motivated by significant commercial and environmental potential, leading industrial companies have launched a wide range of PEV products For example, Tesla Motors pioneered the consumer electric-vehicle market with the Tesla Model S, which costs about $30 per 100 km to drive, versus around $173 per 100 km for a typical premium sedan There are three main categories of electric vehicles, each offering distinct advantages and applications.
• Battery electric vehicle: completely depend on the rechargeable battery.
• Hybrid electric vehicle: combine internal combustion engine with an electric motor and battery The battery is charged by utilizing energy from regenerative breaking.
Plug-in hybrid electric vehicles (PHEVs) use gasoline and an external power outlet to recharge their batteries, allowing them to run on electric power for the first few miles After the battery is depleted, the vehicle automatically switches to hybrid mode, employing both the gasoline engine and electric motor to maintain efficiency.
With the broad uptake of plug-in electric vehicles (PEVs), grid operation faces new challenges around electric vehicle charging The National Electric Code (NEC) classifies PEV charging into three levels—Level 1, Level 2, and DC fast charging (often called Level 3)—each defined by distinct electrical characteristics, wiring requirements, and implications for residential, commercial, and public charging deployments Understanding these levels helps planners and installers design safe, efficient charging infrastructure that supports grid reliability while meeting user charging needs.
• Level-1: Standard 120 V The charging duration is from 6 to 15 h The maximum power falls between 1.44 and 1.92 kW.
• Level-2: 208–240 V The charging duration is 2–5 h, and the maximum charging power is 7.2 kW.
• Level-3: 440/480 V This allows very fast charging, and the charging duration can be 15–30 min.
High PEV penetration levels will lead to overload charging problems for the smart grid For example, electric taxi charging, which is very likely to coincide with
Peak-demand periods threaten distribution feeders with overload, voltage instability, and reduced energy utilization, a challenge intensified by fast PEV charging that demands much higher power than regular charging A large body of research shows that EV charging often coincides with system peak hours, causing grid overload, while smart charging devices can prevent this in a smart grid PEVs also increase demand-side uncertainties and accelerate aging of buses, feeders, and transformers; uncoordinated charging worsens peak load, losses, and voltage drops These challenges, alongside potential benefits such as enhanced frequency regulation, can be effectively addressed through coordinated charging and discharging strategies By implementing coordinated strategies, peak load and losses can be reduced and the adverse effects of uncoordinated charging mitigated Models have examined the optimal and maximum penetration of PEVs in transportation systems (e.g., Ontario), with comparisons also drawn between conventional and hybrid vehicle emissions Overall, coordinated charging/discharging offers a practical approach to balancing demand, stabilizing voltages, and improving energy efficiency while supporting higher PEV adoption.
To date, most studies have emphasized temporal coordination, distributing charging and discharging decisions over different time periods These approaches often assume PEVs are ready for charging or discharging within a defined area, such as parking lots or residential zones In practice, however, PEVs may require fast charging while on the move—such as electric taxis—and mobile vehicles can participate in V2G or V2V transactions if substantial revenue is available Therefore, combining temporal information with spatial coordination for mobile PEV charging and discharging can yield improved system performance, flexibility, and economic returns for grid services.
In the spatial coordination of mobile PEVs, the charging/swapping station assigned for a PEV may be too far from its current location and battery energy level, triggering range anxiety as the travel cost competes with the remaining energy If the PEV battery depletes en route, this can reduce discharging revenues or increase charging costs Drivers therefore favor charging/swapping stations that minimize travel cost, maximize discharging revenue, or minimize charging costs while accounting for range anxiety Yet these preferences can conflict with system technical constraints, underscoring the need for new online charging/discharging strategies that balance drivers’ preferences with operational constraints.
1 In this monograph, the PEV energy consumed on the road to reach a charging/swapping station is referred to as the travel cost.
An Introduction of VANETs
Vehicular ad hoc networks (VANETs) are emerging as a transformative technology that delivers broadband connectivity to moving vehicles by deploying road side units (RSUs) along highways and equipping vehicles with on-board units (OBUs) This framework supports two main communication modes: vehicle-to-RSU (V2R) and vehicle-to-vehicle (V2V) communications, also referred to as vehicle-to-infrastructure (V2I) and inter-vehicle communications, respectively Together, these links enable a wide range of applications, including road-safety services such as accident warnings and traffic alerts, traffic monitoring and management, and infotainment delivery like video streaming and online gaming.
To support automotive applications via VANETs, Vehicle-Infrastructure Integration (VII), also known as IntelliDrive by the U.S Federal Communications Commission (FCC), has allocated a dedicated 75 MHz spectrum in the 5.9 GHz band, now referred to as Direct Short-Range Communications (DSRC) DSRC signals can reach about 1 km under permitted power levels, with data rates ranging from 6 to 27 Mbps The standardization effort for DSRC covers the technical details of the PHY and MAC layers and the overall communication architecture, which envisions ad hoc communications between vehicle-based OBUs and roadside units (RSUs) In VANETs, RSUs function as data repositories or repeaters, enabling safety applications that warn drivers about potentially conflicting situations based on information received from neighboring vehicles or RSUs directly.
VANETs offer enormous potential for safety applications in transportation networks On February 3, 2014, the U.S Department of Transportation's National Highway Traffic Safety Administration announced plans to enable vehicle-to-vehicle (V2V) communications, allowing cars to exchange basic safety data to prevent crashes In parallel, recognizing strong commercial potential, pioneering industrial companies have launched multiple projects to advance vehicular communications and deploy V2V and related ITS technologies.
“Toyota Friend” established a private mobile social network for the Toyota car owners [18].
The vehicular environment presents unique opportunities, challenges, and requirements for vehicle communications High vehicle velocity and rapidly changing channel conditions create significant transmission challenges In highly mobile environments, this complexity calls for an entirely new paradigm for vehicle communications.
2 DSRC protocol supports both RSU-to-vehicle/vehicle-to-RSU (R2V/V2R) and vehicle-to-vehicle(V2V) communication.
6 1 Introduction safety applications can be established, and even other non-safety applications can significantly improve the road and vehicle efficiency.
Applications of VANETs are generally categorized into two broad groups: safety-related applications that improve road safety for drivers and passengers, and non-safety applications that deliver value-added services like vehicle navigation and path planning through vehicular networks.
Inter-vehicular safety applications are designed to alert drivers to potential hazards in real time by sharing current positions, speeds, and accelerations of nearby vehicles This approach is increasingly practical, as most consumer vehicles are already equipped with sensors to measure velocity and acceleration and with transceivers that support vehicle-to-vehicle (V2V) and vehicle-to-roadside (V2R) communications This data-sharing enables timely warnings and safer driving on the move.
Non-safety related applications, such as comfort-driven driving apps, rely on cooperative data collection and sharing across large-scale road networks, including highway and urban road corridors A typical example is VANET-based sensing data-sharing systems, which provide distributed, citywide traffic-condition information This shared data can be used to regulate vehicle traffic flows and reduce overall travel time across the network.
VANETs (Vehicular Ad Hoc Networks) are a specialized form of mobile ad hoc networks (MANETs) tailored for automotive environments In VANETs, vehicles equipped with onboard units (OBUs) communicate with other vehicles and with fixed infrastructure such as roadside units (RSUs), as illustrated in Fig 1.2 This communication framework enables rapid data exchange and cooperative applications on the road Compared with general MANETs, VANETs exhibit distinctive characteristics driven by high mobility and dynamic topology, along with the role of infrastructure-assisted connectivity.
In VANETs, topology changes are rapid due to high vehicle mobility, making the network highly dynamic Vehicle trajectories typically mirror the road geometry, following freeways and the street patterns that define real-world traffic routes.
In VANETs, vehicular communications are not bound by conventional power constraints because vehicle batteries are rechargeable, enabling sustained operation This gives VANETs a distinct advantage over MANETs, where handheld devices face energy limitations that can restrict network longevity.
3 Large scale:A VANET consist of a large number of vehicles, with the scale of the number of vehicles approximately10 7 in reality [19].
Variable network density describes how vehicle counts vary across both time and space Traffic is not uniform; density spikes along downtown corridors during rush hours and remains lower in other areas and times Understanding this temporal and spatial variability is essential for predicting travel times, optimizing routes, and improving transportation system resilience.
High predictable mobility in urban transportation is shaped by road patterns, with urban vehicle speeds typically ranging from 0 to 60 km/h and average highway speeds reaching up to 100 km/h, so mobility is largely regulated by road design and traffic patterns.
Most VANET applications rely on connections to road-side units (RSUs) to communicate with remote servers, and these links depend on multi-hop V2V relaying and V2R communications While VANETs promise a bright future for connected mobility, achieving efficient vehicular communications remains hampered by fundamental challenges.
Network connectivity is the primary challenge in VANETs because high vehicle mobility causes rapid topology changes and intermittent links Many Internet-based VANET applications, such as vehicular video conferencing and traffic monitoring, rely on stable connections to remote servers via roadside units (RSUs) To extend the limited range of vehicle-to-roadside (V2R) links, inter-vehicle relaying through vehicle-to-vehicle (V2V) communications is commonly employed In an uplink scenario, three vehicles can collaboratively relay data toward RSUs, which then forward it to the remote server However, the fast-changing topology due to vehicle movement leads to transient connections between vehicles, resulting in highly unreliable inter-vehicle transmissions.
One of the central challenges in VANETs is the large scale of the vehicular network, where application performance depends on how many vehicles contend for transmissions and how available RSUs are Analyzing how nodal throughput scales with the vehicle population and RSU deployment patterns—capturing the asymptotic network throughput capacity—helps in selecting appropriate network mechanisms, such as MAC protocols and relay-selection schemes, and in guiding practical RSU deployment This throughput scaling insight supports network planning and mechanism design to sustain reliable vehicular communications as the network grows.
Architecture of VANET-Enhanced Smart Grid
The Heterogeneous Wireless Network
In the literature, most studies exploit cellular networks (GSM, UMTS, LTE) to support PEV charging and discharging strategies Cellular networks offer significant advantages for wireless transmission because base stations provide wide coverage and reliable connectivity This communication framework enables real-time data exchange for control, pricing signals, and coordination between PEVs and the electric grid The expansive base-station coverage makes cellular networks a scalable and robust option for managing PEV energy flows while minimizing grid impact Overall, cellular-enabled communication remains a preferred approach due to its maturity, interoperability, and extensive existing infrastructure.
1.4 Architecture of VANET-Enhanced Smart Grid 9
Cellular networks have inherent drawbacks that limit their effectiveness in collecting real-time vehicle information Since cellular systems are not designed specifically for vehicular communications, data delivery can be costly and bandwidth-intensive In addition, the large volume of vehicular data can congest other cellular services, especially in areas with high vehicle density.
VANETs have recently emerged as a promising technology for revolutionary wireless broadband communications for vehicles By deploying roadside gateways along highways and sidewalks (RSUs) and equipping vehicles with on-board units (OBUs), VANETs enable vehicle-to-RSU (V2R) and vehicle-to-vehicle (V2V) communications, supporting real-time information exchange as vehicles move Designed for multi-hop information exchange among highly mobile vehicles and RSUs, VANETs can deliver real-time data efficiently via short-range V2V and V2R links, making large-volume vehicle information collection and dissemination cheaper than cellular networks However, high mobility and short-range transmission cause intermittent connections between vehicles and RSUs, leading to considerable transmission delay for real-time information delivery This delay can affect the effectiveness of PEV charging/discharging decisions, since moving PEVs continue to consume energy while waiting Consequently, the additional travel cost induced by VANET transmission delay should be explicitly considered.
Integrating cellular systems with VANETs into a heterogeneous wireless network combines their advantages and mitigates their limitations, enabling more efficient information delivery through a low-cost, low-delay communication framework The following sections discuss how this heterogeneous network can deliver real-time messages and support the design of online charging/discharging strategies.
Heterogeneous Wireless Network-Enhanced
Figure 1.3 illustrates the proposed heterogeneous wireless-network-enhanced smart grid architecture, integrating a power distribution system with distributed generators (DGs), charging stations, swapping stations, plug-in electric vehicles (PEVs), road-side communication access points (RSUs) along roadways, and the base stations (BSs) of cellular networks.
4 Note that, in this monograph, the term V2V is used in two different contexts, one for VANET communications among vehicles and the other for energy transfer among PEVs.
: Wired connected : V2R/R2V communication : V2V communication : a charging station
: Wireless communication via a cellular network
: Power line Fig 1.3 Heterogeneous wireless network-enhanced smart grid
The power distribution system consists of a substation and a set of distributed generators (DGs) that deliver energy across the network through power feeders to buses Charging stations located at different buses provide fast-charging services for plug-in electric vehicles (PEVs) via grid-to-vehicle (G2V) power transfer PEVs can also feed electricity back into the grid through vehicle-to-grid (V2G) mode when charging at these stations In addition, energy can be exchanged at swapping stations, enabling PEVs to transfer energy without connecting to the power grid via vehicle-to-vehicle (V2V) charging and discharging.
The time horizon is partitioned into consecutive periods of fixed duration At the start of each period, the electricity price is determined from the collected PEV charging information and the system load capacities Based on historical RTU readings, the model also predicts the maximal power that can be exchanged at the swapping stations in the following period, i.e., the load-capacity of the swapping stations In this context, denote Bus_j as the j-th bus at B_j, and let C_j represent the load-capacity of the charging/swapping station at B_j.
A set of roadside RSUs is deployed along the roads to collect PEV charging and discharging information through V2R transmissions Cellular base stations provide the wireless communication links between BSs and the portable transceivers in PEVs Both RSUs and BSs are connected to charging stations via wired lines and can relay the collected PEV decisions to the charging stations for price updates The pricing is updated to balance the system load by aligning prices with the observed load capacities and demand.
Aim of This Monograph
To handle demand fluctuations, roadside units (RSUs) and base stations (BSs) retrieve the latest pricing from charging and swapping stations, and then relay this updated price information to plug-in electric vehicles (PEVs) through R2V (roadside-to-vehicle) and V2V (vehicle-to-vehicle) communications or via the cellular network, enabling real-time price awareness for optimal charging or swapping decisions.
Let V denote the set of mobile plug-in electric vehicles (PEVs) Each PEV is equipped with both cellular and VANET interfaces, enabling real-time information exchange Depending on the scenario, PEV data can be disseminated via multi-hop V2V relaying and V2R transmission or delivered to a cellular base station through portable transceivers Upon receiving control messages, such as price control instructions, individual PEVs determine their charging or discharging actions while accounting for range anxiety The charging or discharging decision for PEV v at bus B_j during period k is denoted P_v;j;k, together with an indicator that specifies whether PEV v will head to the charging or swapping station at bus B_j.
B j in periodkfor charging/discharging (denoted asx v; j ; k ) The variablex v; j ; k is set to
At station B_j in period k, the decision variable PEV_v;j;k indicates whether a plug-in electric vehicle (PEV) will be charged or discharged, with PEV_v;j;k > 0 for charging, PEV_v;j;k < 0 for discharging, and PEV_v;j;k = 0 otherwise If the PEV chooses to charge, the corresponding charging load P_v;j;k is positive; if it chooses to discharge, the charging load becomes negative After all PEVs have made their charging or discharging decisions, these actions are communicated back to the charging and swapping stations via a heterogeneous wireless network.
In this monograph, we focus on leveraging real-time vehicle information to design an efficient online PEV charging and discharging strategy with joint spatial and temporal coordination We investigate three core problems: (1) what information is required to support a spatially and temporally coordinated online charging/discharging framework; (2) how to efficiently and reliably obtain the real-time information needed for the PEV online charging/discharging strategy; and (3) based on the collected real-time information, how to design the mobility-aware coordinated PEV charging/discharging strategy to improve overall power utilization and reduce charging/discharging costs.
This monograph examines a VANET-enhanced smart grid for fast charging of plug-in electric vehicles (PEVs), leveraging VANETs and VANET-involved heterogeneous networks to collect vehicle mobility information and dispatch charging and discharging decisions in real time It proposes a mobility-aware coordinated fast-charging strategy for EVs that improves overall energy utilization, prevents power-system overloading, and reduces EV range anxiety by lowering the average travel cost To further boost charging efficiency in a smart grid, a vehicle-to-vehicle energy swapping scheme is also introduced to offload heavy charging demand from the grid.
Shifting the EV charging load from the smart grid to the swapping station, this approach aims to relieve the grid's charging burden by incentivizing EVs with surplus energy to participate in EV charging, and to maximize revenues from discharging EVs while minimizing the cost of charging EVs.
1 A comprehensive guide to plug-in hybrid vehicles, Hybrid Cars (2011), [Online] available: http://www.hybridcars.com/plug-in-hybrid-cars/#battery
2 Electric Power Resech Institute, [Online] available: http://www.epri.com/Pages/Default.aspx
3 C Liu, K Chau, D Wu, S Gao, Opportunities and challenges of vehicle-to-home, vehicle-to- vehicle, and vehicle-to-grid technologies Proc IEEE 101(11), 2409–2427 (2013)
4 TESLA Motors, [Online] available: http://www.teslamotors.com/Pages/goelectric#
5 P Richardson, D Flynn, A Keane, Local versus centralized charging strategies for electric vehicles in low voltage distribution systems IEEE Trans Smart Grid 3(2), 1020–1028 (2012)
6 G.T Heydt, The impact of electric vehicle deployment on load management strategies IEEE Trans Power Appar Syst 1(144), 1253–1259 (1983)
7 A Heider, H.J Haubrich, Impact of wide-scale EV charging on the power supply network, in IEEE Colloquium on Electric Vehicles – A Technology Roadmap for the Future, London, vol 6, no 262, 1998, pp 1–4
8 K Schneider, C Gerkensmeyer, M Kintner-Meyer, M Fletcher, Impact assessment of plug-in hybrid electric vehicles on Pacific Northwest distribution systems, in IEEE Power and Energy Society 2008 General Meeting, Pittsburgh, US 2008
9 C Roe, F Evangelos, J Meisel, S Meliopoulos, T Overbye, Power system level impacts of PHEVs, in Proceedings of the 42nd Hawaii International Conference on System Sciences,
10 K Clement-Nyns, E Haesen, J Driesen, The impact of charging plug-in hybrid electric vehicles on a residential distribution grid IEEE Trans Power Syst 25(1), 371–380 (2010)
11 K Clement, E Haesen, J Driesen, Coordinated charging of multiple plug-in hybrid electric vehicles in residential distribution grids, in Proceedings of the Power Systems Conference and Exposition, Seattle, US 2009
12 A Hajimiragha, C.A Canzares, M.W Fowler, A Elkamel, Optimal transition to plug-in hybrid electric vehicles in Ontario, Canada, considering the electricity-grid limitations IEEE Trans. Ind Electron 57(2), 690–701 (2010)
13 C.H Stephan, J Sullivan, Environmental and energy implications of plug-in hybrid-electric vehicles Environ Sci Technol 42(4), 1185–1190 (2008)
14 L Cheng, B.E Henty, D.D Stancil, F Bai, P Mudalige, Mobile vehicle-to-vehicle narrow- band channel measurement and characterization of the 5.9 GHz dedicated short range communication (DSRC) frequency band IEEE J Sel Areas Commun 25(8), 1501–1516
15 H.T Cheng, H Shan, W Zhuang, Infotainment and road safety service support in vehicular networking: from a communication perspective Mech Syst Signal Process 25(6), 2020–2038
16 Vehicle Infrastructure Integration (VII) Program, http://www.vehicle-infrastructure.org/
18 Toyota Friend by Toyota, http://pressroom.toyota.com/releases/toyota+friend+social+network. htm
19 Y Khamayseh, M Yassein, M Alghani, C Mavromoustakis, Network size estimation inVANETs Netw Protoc Algorithms 5(3), 136–152 (2013)
20 M Abboud, L Jaoude, Z Kerbage, Real time GPS navigation system (2004), http://webfea- lb.fea.aub.edu.lb/proceedings/2004/SRC-ECE-27.pdf
21 M Grossglauser, D Tse, Mobility increase the capacity of ad hoc wireless networks. IEEE/ACM Trans Netw 10(4), 477–486 (2002)
22 L Gan, U Topcu, S Low, Optimal decentralized protocol for electric vehicle charging IEEETrans Power Syst 28(2), 940–951 (2013)
Several studies show that high penetration of plug-in electric vehicle (PEV) charging can significantly impact the power system, while coordinated discharging offers substantial benefits to grid reliability and efficiency To efficiently implement these principles, key challenges must be addressed, including PEV mobility modeling, transmission network selection, balancing the power system’s technical limitations with drivers’ preferences, and establishing viable business revenue models for V2G and V2V transactions.
Classifications of Charging/Discharging Strategies
The research works on charging/discharging strategy design for PEVs in the smart grid can be categorized from different perspectives as follows:
This article analyzes centralized versus decentralized charging and discharging strategies for plug-in electric vehicles (PEVs) In centralized control, a central controller coordinates charging/discharging to reach globally optimal solutions, but this approach entails high signaling overhead for information collection and significant computation requirements By contrast, decentralized strategies let PEVs make local, iterative decisions, reducing computation at the expense of slower convergence and potentially suboptimal outcomes due to limited information exchange In this monograph, we exploit VANETs for real-time information delivery and a powerful remote traffic server to implement a centralized charging/discharging strategy that achieves globally optimal performance, with the detailed methodology and results presented in Chapter 3.
• Pure PEV coordination (e.g., [1]) and price control-based strategy (e.g., [2]):
Pure PEV coordination assumes that drivers unconditionally follow charging decisions However, for mobile PEVs, those decisions can conflict with the system’s technical limitations and the drivers’ preferences.
On the contrary, price control based strategy can effectively address such a © Springer International Publishing 2016
M Wang et al., Mobile Electric Vehicles, Wireless Networks,
The EV charging and discharging problem is driven by electricity prices, which charging stations set jointly based on the anticipated demand and supply from PEV charging and discharging activities This monograph proposes a price-control strategy described in Chapter 4 that incentivizes drivers to follow a coordinated plan by offering lower charging prices and higher revenue for discharging.
Two approaches—myopic charging/discharging strategies that rely only on current grid information, and predictive strategies that also incorporate future power demands—shape how PEV charging/discharging interacts with the grid Since predictive strategies account for both present and anticipated loads through short-term power-load forecasting, their decisions are more reliable in real-world operation Therefore, this monograph proposes a predictive charging/discharging strategy in Chap.3.
Electric Vehicle Charging Strategy Design
Extensive studies show that high EV charging penetration can markedly impact power systems To prevent overload during peak hours, load management strategies are needed to shift charging demand across time and space Peak load shifting to off-peak periods can improve system load factors and reduce the risk of overload [1,8] Comparative studies [9,10] indicate that global charging strategies that coordinate charging duration and rates based on overall load information outperform local approaches in energy utilization Incorporating the spatial diversity of charging [11,12] further helps shape the charging profile Yet most existing strategies assume EVs are stationary during charging; mobility is often neglected, even though it is a fundamental vehicle characteristic, especially for fast charging Mobility introduces range anxiety—the tension between travel cost and battery level—affecting charging feasibility Consequently, there is a need for new, efficient EV charging strategies that leverage real-time mobility data to mitigate range anxiety and optimize charging decisions.
Most current approaches to obtaining real-time vehicular information rely on cellular or Wi‑Fi networks However, these systems have inherent drawbacks that limit their practicality for collecting vehicle data In dense vehicular networks, the location measurement inaccuracies of these systems can significantly degrade charging performance Moreover, since cellular networks are not exclusive to vehicular communications, data collection can be costly, and the large volumes of vehicular data transmission may lead to scalability and cost challenges.
Challenging Issues for Charging/Discharging Strategy Design
Mobility Modeling of PEVs
PEV mobility directly influences the travel cost to charging or swapping stations and, in turn, shapes the coordinated charging and discharging decisions of the fleet In a suburban setting, PEVs are assumed to follow a mobility pattern described by established models such as Wiedemann 74 The mobility of each PEV can be characterized by random variables; let S denote the vehicle velocity, which can take a finite set of values When there are two velocity states (n = 2), S assumes the low velocity S_L and the high velocity S_H The transitions between these two states are described by a two-state continuous-time Markov chain with state transition rates LH (low-to-high) and HL (high-to-low).
The proposed model can be leveraged to faithfully depict real-world human driving behaviors, showing how drivers typically maintain a given velocity for a period and then switch to another speed based on personal choice, road conditions, or the headway distance to the vehicle ahead By incorporating a continuous state representation, the model captures smooth transitions between speeds and driving regimes, enabling more accurate simulations of traffic dynamics under varying conditions This approach supports realistic traffic analysis and optimization by reflecting how drivers adapt their speed in response to dynamic road and vehicle interactions.
Markov chain of the Wiedemann 74 model, the headway distance between two
Section 18.2 derives key intermediate metrics for electric-vehicle networks, including charging and discharging interactions among neighboring EVs and plug-in electric vehicles (PEVs) within a single lane, the inter-contact time between PEVs, and overall vehicle density These results serve as inputs for calculating travel cost and the energy consumed due to transmission delay, providing a foundation for assessing efficiency in charging strategies, routing decisions, and the impact of communication latency on energy use.
With the mobility model and the calculated travel cost and transmission delay,the range anxiety can be specified.
Network Selection for Real-Time Information Delivery
In VANETs, high PEV mobility and the short-range nature of V2V and V2R links make connections intermittent, which introduces transmission delays and an extra travel distance (cost) as PEVs wait for charging decisions Using cellular networks to deliver information can incur additional monetary costs To address this, a network selection mechanism is needed to balance the tradeoff between the travel cost resulting from VANET transmission delays and the monetary cost of cellular connectivity, thereby optimizing charging decisions and overall system performance.
Let d_v;k denote the travel distance of a Plug-in Electric Vehicle (PEV) during period k while waiting for a decision from Vehicular Ad-hoc Networks (VANETs) Based on the PEV mobility model described earlier, this distance can be expressed as a function that incorporates the impact of VANET transmission delay on travel, linking d_v;k to the delay experienced in VANET communication The transmission delay in VANETs depends on network conditions and traffic, and it decreases as network performance improves, thereby shaping the travel distance observed during the waiting period.
1 the vehicle velocitySand the inverse of the parameter 1 , since increasingSand
1 will reduce the average number of hops in a multi-hop transmission link, thus leading to a reduced transmission delay,
2 the vehicle density (i.e.,), since increasing the vehicle density will create more chances for a successful transmission, thus potentially reducing the average transmission delay for a multi-hop communication link,
3 higher possibility for V2R transmissions in the network (i.e., larger trans- mission range R or more deployed RSUs to decrease the average inter-RSU distanceL), and
4 a more efficient transmission mechanism, e.g., choosing the farthest vehicle within its transmission range as the relay to reduce the potential number of transmission hops, or designing a more efficient MAC protocol to avoid transmission collisions among multiple transmission pairs.
In VANETs, the travel cost caused by transmission delay is modeled as a linear, non-decreasing function PC(d_v, k) that quantifies the energy cost for a PEV v to wait for a decision in period k This cost function ties communication latency directly to the vehicle's energy expenditure, enabling more accurate evaluation of routing and scheduling decisions Since PC(d_v, k) increases linearly with the delay d_v, longer delays lead to higher energy costs for the PEV, guiding optimization strategies to minimize latency and conserve energy.
By balancing the VANET travel costP v; cost k with the cellular network monetary cost, the dynamic and adaptive network selection can make the information delivery more efficient and economic.
Balancing the Tradeoff Between the Power
Technical Limitations and Drivers’ Preferences
Drivers typically choose a charging or swapping station based on personal preferences, such as the shortest route length or the station they're most familiar with However, there are times when the preferred station cannot support additional power loads To prevent system overload, another charging or swapping station is assigned to the PEV for charging or discharging, a process known as spatial coordination.
Travel cost for a plug-in electric vehicle (PEV) v in period k, measured in energy terms and denoted as PC(dv,k), should not exceed the discharging revenue; alternatively, setting the electricity price sufficiently low can motivate drivers to visit the designated charging or swapping station The travel cost can also be formulated based on other driver preferences, for a subset of stations along the customer's route For example, if a driver’s route is known, the driver will prefer to select a charging station along that route, meaning the candidate charging stations should be limited to those deployed along the route This route-specific subset of charging stations can be incorporated into the optimization problem as an additional constraint.
Drivers' preferences for when and how to charge or discharge their electric vehicles create a tradeoff between optimal power-system utilization and individual choice This tension highlights two main challenges: defining each driver's charging and discharging preferences and balancing the power grid's technical limitations with those preferences Addressing these challenges requires accurate modeling of driver preferences and robust optimization that respects grid constraints while meeting user needs Deploying this balance through targeted control strategies and incentive mechanisms can improve grid efficiency, reliability, and driver satisfaction without sacrificing system performance.
1 P Richardson, D Flynn, A Keane, Local versus centralized charging strategies for electric vehicles in low voltage distribution systems IEEE Trans Smart Grid 3(2), 1020–1028 (2012)
2 L Gan, U Topcu, S Low, Optimal decentralized protocol for electric vehicle charging IEEE Trans Power Syst 28(2), 940–951 (2013)
3 Y Cao, S Tang, C Li, P Zhang, Y Tan, Z Zhang, J Li, An optimized EV charging model considering TOU price and SOC curve IEEE Trans Smart Grid 3(1), 388–393 (2012)
4 M F Shaaban, M Ismail, E.F El-Saadany, W Zhuang, Real-time PEV charging/discharging coordination in smart distribution systems IEEE Trans Smart Grid 5(4), 1797–1807 (2014)
5 M Shaaban, Y Atwa, E El-Saadany, PEVs modeling and impacts mitigation in distribution networks IEEE Trans Power Syst 28(2), 1122–1131 (2013)
6 D Ban, G Michailidis, M Devetsikiotis, Demand response control for PHEV charging stations by dynamic price adjustments, in Proceedings of the IEEE Innovative Smart Grid Technologies,
7 R.C Green, L Wang, M Alam, The impact of plug-in hybrid electric vehicles on distribution networks: a review and outlook J Renew Sustain Energy Rev 15(1), 544–553 (2011)
8 G.T Heydt, The impact of electric vehicle deployment on load management strategies IEEETrans Power Appar Syst 1(144), 1253–1259 (1983)
9 K Mets, T Verschueren, W Haerick, C Develder, F Turck, Optimizing smart energy control strategies for plug-in hybrid electric vehicle charging, in Proceedings of the IEEE Network Operations and Management Symposium Workshops, Osaka, Apr 2010
10 K Clement, E Haesen, J Driesen, Coordinated charging of multiple plug-in hybrid electric vehicles in residential distribution grids, in Proceedings of the IEEE Power Systems Conference and Exposition, Seattle, Mar 2009
11 L Kelly, Probabilistic modeling of plug-in hybrid electric vehicle impacts on distribution networks in British Columbia, M S thesis, Department of Mechanical Engineering, University of Victoria, Victoria, 2009
12 S Bae, A Kwasinski, Spatial and temporal model of electric vehicle charging demand IEEE Trans Smart Grid 3(1), 394–403 (2012)
13 J Herrera, D Work, R Herring, X Ban, A Bayen, Evaluation of traffic data obtained via GPS-enabled mobile phones: the mobility century field experiment, in Working Paper UCB- ITS-VWP-2009-8, Aug 2009
14 R Herring, A Hofleitner, S Amin, Using mobile phones to forecast arterial traffic through statistical learning, in Proceedings of the 89th Annual Meeting of Transportation Research Board, Washington, DC, Jan 2010
15 K Lee, J Lee, Y Yi, I Rhee, S Chong, Mobile data offloading: how much can WiFi deliver? in Proceedings of the ACM Co-NEXT, New York, Nov 2010
16 H Liu, A Danczyk, R Brewer, R Starr, Evaluation of cell phone traffic data in Minnesota. Transp Res Rec 2086(1), 1–7 (2008)
17 N Lu, N Zhang, N Cheng, X Shen, J Mark, F Bai, Vehicles meet infrastructure: toward capacity-cost tradeoffs for vehicular access networks IEEE Trans Intell Transp Syst 14(3),
18 M Wang, H Shan, L.X Cai, N Lu, X Shen, F Bai, Throughput capacity of VANETs by exploiting mobility diversity, in Proceedings of the IEEE ICC’12, Ottawa, June 2012
19 N Lu, T Luan, M Wang, X Shen, F Bai, Bounds of asymptotic performance limits of social proximity vehicular networks IEEE/ACM Trans Netw 22(3), 812–825 (2014)
20 R Wiedemann, Modeling of RTI-elements on multi-lane roads, in Proceedings of the Drive Conference, Brussels, Feb 1991
Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid
Coordinated charging enables efficient EV charging and better energy utilization while preventing overload of the power system Designing such a strategy is challenging because it must steer mobile EVs toward fast-charging stations in a globally optimal way This work analyzes a VANET-enhanced smart grid that uses vehicular ad-hoc networks to support real-time communications among mobile EVs and between EVs and roadside units (RSUs) for timely mobility data and charging decisions We introduce a mobility-aware coordinated charging strategy that improves overall energy utilization, safeguards the power system, and reduces range anxiety by managing average travel costs Specifically, we account for mobility-related travel costs in two aspects.
(1) the travel distance from the current EV location to the fast-charging station, and
(2) the transmission delay for an EV to receive a charging decision through theVANETs.
Introduction
Electric vehicles (EVs) are increasingly recognized as a cornerstone of sustainable and eco-friendly transportation By running on electricity instead of gasoline, EVs can deliver substantial savings over a vehicle’s life, with figures showing the Tesla Model S costing about $30 per 100 km to drive compared with around $173 per 100 km for a typical premium gasoline sedan Widespread EV adoption also lowers overall energy use in the transport sector and reduces emissions, advancing greener mobility worldwide.
M Wang et al., Mobile Electric Vehicles, Wireless Networks,
Mobility-aware coordinated EV charging in a VANET-enhanced smart grid integrates conventional energy sources with electric vehicles to lower environmental pollution and carbon footprints Industry reports indicate that battery electric vehicles, powered entirely by rechargeable batteries and producing zero tailpipe emissions, can reduce overall transport-sector emissions by as much as 70%, fueling a higher market share for EVs According to the Electric Power Research Institute (EPRI), projections for 2020 and 2030 point to continued growth in EV adoption and its impact on the transport sector.
By 2050, EV penetration is projected to reach 35%, 51%, and 62% in different scenarios However, widespread EV adoption could lead to charging coinciding with peak power demand, risking overload on distribution feeders, system instability, and lower overall energy utilization—especially for fast charging, which demands much higher power than ordinary charging To mitigate the impact on the power system without extra deployment costs, some studies propose energy storage systems, though they add cost Alternatively, load management strategies that coordinate EV charging in time and space can flatten peaks and boost energy utilization At the same time, for fast EV charging, designated charging stations must be reachable by mobile EVs given their current locations and battery levels to address range anxiety.
Extensive literature on coordinated EV charging has proposed many strategies, but most studies address only the power-system perspective within a localized area where a group of EVs is assumed ready to charge Few works incorporate vehicle mobility and the fast-charging context into charging strategies In practice, EVs may need to be charged while traveling, so the travel cost to reach a charging station must be considered Without accounting for travel cost and current location and battery level, an EV may fail to reach the assigned charger under existing strategies Range anxiety drives drivers to prefer charging locations with lower travel cost, underscoring the need for charging strategies that model mobility and travel cost to effectively reduce travel burden Real-time collection of EV data, including locations and battery levels, is essential to support informed charging decisions.
This chapter focuses on leveraging real-time EV mobility data to design a coordinated, smart charging strategy that is both efficient and scalable By integrating current transportation patterns with grid analytics, the approach aligns charging sessions with low-demand periods and higher renewable output, boosting overall energy utilization while reducing the average charging cost, grid stress, and vehicle downtime This framework supports demand-side management and enables more reliable, greener, and cost-effective electric vehicle charging for a coordinated future.
To manage EV travel costs and prevent overload of the power system, a well-designed charging strategy must address two core questions: first, how to efficiently and reliably acquire real-time information from mobile vehicles required by the EV charging strategy; and second, how to use this mobility data to enable mobility-aware, grid-friendly optimization By integrating accurate vehicle location, state of charge, and travel patterns, the charging system can anticipate demand, schedule charging during favorable grid conditions, and distribute load to avoid peaks, thereby lowering total energy costs and enhancing grid stability This data-driven approach supports resilient, cost-effective EV adoption and smoother grid operation, guiding charging timing, location, and capacity planning for optimal performance.
3.1 Introduction 23 coordinated EV charging to improve energy utilization and reduce EV travel cost while preventing power system from overload.
Leveraging vehicular ad-hoc networks (VANETs) offers a promising way to address the initial problem VANETs are designed for multi-hop information exchange among highly mobile vehicles and roadside units (RSUs), with short-range vehicle-to-vehicle (V2V) and vehicle-to-RSU (V2R) communications enabling rapid data dissemination across dynamic traffic environments [21].
VANETs-enabled real-time information delivery enables efficient, large-scale collection of vehicle data at a lower cost and faster speed than cellular networks or Wi-Fi Roadside units (RSUs) significantly enhance the timeliness of data collection and dissemination, making coordinated charging strategies for groups of mobile EVs feasible Consequently, this chapter integrates VANETs into the smart grid to collect real-time information from moving electric vehicles and to inform charging decisions Since energy is still consumed while vehicles wait for charging decisions, the information-exchange delay in VANETs can incur additional travel costs; therefore, we analyze transmission delay as a function of vehicle density and RSU deployment.
To address range anxiety in electric-vehicle (EV) charging, this work proposes a mobility-aware coordinated charging strategy that uses historical remote terminal unit (RTU) readings and real-time vehicle information to guide decisions It answers three key questions: given a vehicle’s current state of charge, should it be charged in the next period; which charging station should serve the vehicle considering its location and range concerns; and how much energy should be charged to improve overall energy utilization while ensuring power-system stability The optimal charging problem is formulated as a time-coupled mixed-integer linear program (MILP), which is computationally challenging, but by uncovering linear relationships among EV charging loads on feeders, the MILP is time-decoupled into a set of sub-MILPs via Lagrange duality Each sub-MILP can then be solved with a branch-and-cut-based outer-approximation algorithm.
To address EV range anxiety, this work integrates VANETs with the smart grid to collect real-time vehicle data for mobility tracking, including locations and battery levels Building on this data, it proposes a predictive, mobility-aware, coordinated EV charging strategy that aims to optimize power utilization, reduce average EV travel costs, and prevent power-system overload in the next charging period The chapter makes four key contributions: a VANET-smart grid data fusion framework for accurate mobility and energy state estimation; a predictive, mobility-aware charging algorithm that coordinates charging across the network; improved power efficiency and lower travel costs for EV users; and enhanced grid reliability by mitigating peak loads in upcoming charging windows.
We propose a VANET-enhanced smart grid architecture in which VANETs enable efficient communication among mobile EVs and roadside units (RSUs) to collect relevant information and disseminate EV charging decisions in real time The gathered data are processed by a traffic server, which implements a predictive, coordinated EV charging strategy to optimize charging schedules and improve grid performance.
24 3 Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid
• Second, considering the range anxieties of EVs, a mobility-aware coordinated
An EV charging strategy is designed to enhance the overall energy utilization of the power system, reduce the average EV travel cost, and prevent overload of the grid It reveals a linear relationship among the charging loads at charging stations, a pattern essential for effective load distribution In addition, the travel cost is defined and formulated to demonstrate how charging behavior influences costs and system performance.
EV mobility impacts the charging decision making;
The globally optimal charging problem is formulated as a time-coupled mixed-integer linear program (MILP) By exploiting Lagrange duality, this formulation can be decoupled into a set of sub-MILPs, enabling scalable and parallelizable solution Each sub-MILP is then solved using a branch-and-cut–based outer-approximation algorithm, which efficiently tightens relaxations to arrive at the optimal charging plan.
Extensive simulations validate the efficacy and efficiency of the proposed EV charging strategy VISSIM is used to extract simulation traces, enabling the construction of a highly realistic suburban scenario The study thoroughly evaluates the transmission delay induced by VANETs within this context Results show that the proposed strategy significantly outperforms traditional autonomous charging approaches that operate without VANETs, in terms of energy utilization and the average EV travel cost.
The remainder of this chapter is organized as follows Section3.2 elaborates the system model Sections3.3and3.4present the formulation and solution of the mobility-aware coordinated EV charging problem, respectively The performance of the proposed strategy is demonstrated in Sect.3.5via simulations The related works are introduced in Sect.3.6 Section3.7concludes this chapter.
System Model
VANET-Enhanced Smart Grid
Figure 3.1 gives the components of the proposed VANET-enhanced smart grid architecture The architecture consists of a power distribution system, charging stations (e.g., at parking lots), a traffic server, RSUs along the road sides and EVs.Energy to the whole network is supplied by the power distribution system via power feeders (i.e., buses) The charging stations provide fast-charging for all the EVs.
: Wired connected V2R/R2V communication V2V communication : a charging station
: Wireless communication via a cellular network
: Power line Fig 3.1 The VANET-enhanced smart grid architecture
Using historical RTU readings from each bus in the distribution system, the voltage at every charging station in the next period can be forecast, allowing the calculation of each station’s load-capacity, denoted as P_j^total for Bus j The RTU data are transmitted to the traffic server over wired links, where a predictive charging strategy is executed to determine globally optimal charging decisions for the EVs in need of charging, based on real-time EV information from VANETs and the historical RTU measurements The strategy operates period by period; for period k, the charging load/rate of EV v at Bus j is P_ch^v;j;k, and the charging indicator x_v;j;k specifies whether EV v should be charged at station j in period k, with x_v;j;k = 1 if scheduled for charging at Bus j in period k, and 0 otherwise.
In VANETs, a set of electric vehicles (EVs) moves through a network region along map-based paths, with V denoting the EV set EVs may need charging while traversing roads, highlighting the importance of real-time energy-aware coordination Real-time EV information—covering location, energy level, and charging needs—can be exchanged among on-board units (OBUs) via multi-hop V2V relaying, enabling cooperative routing, charging coordination, and data sharing across the roadway network.
Mobility-aware coordinated EV charging in a VANET-enhanced smart grid is implemented under the DSRC protocol with transmission range R EVs are equipped with GPS devices for shortest-path navigation and wirelessly connected to their onboard units (OBU) A set of RSUs is uniformly deployed along roadways to collect EV information—such as positions and battery levels—via DSRC-based V2R transmissions with range R These RSUs relay the collected data to a traffic server through wireline connections, enabling globally optimal EV charging decisions When the RSUs obtain the EV information, they forward it to the traffic server to support coordinated charging optimization across the network.
EV charging decisions from the traffic server, they will relay the decisions back to the EVs through R2V and V2V transmissions.
To summarize, the VANET-enhanced smart grid system operates as follows.
Information collection and reporting to the traffic server is two-fold: first, the historic RTU readings of each bus in the power system are conveyed to the traffic server via wireline, enabling estimation of the charging load constraints at each charging station; second, the real-time EV information is gathered through multi-hop V2V relaying and V2R transmissions.
Predictive coordinated EV charging uses a traffic server to fuse gathered information from traffic flows, grid conditions, and charging demand, enabling optimal EV charging decisions This data-driven approach maximizes the grid’s power utilization, reduces the average EV travel cost, and prevents power system overloading by balancing supply and demand in real time.
During the decision dissemination phase, the EV's onboard unit (OBU) receives its charging decision from nearby vehicles or from the road-side unit (RSU) and immediately forwards this decision to the GPS navigation system The GPS then guides the vehicle to the assigned charging station, ensuring efficient routing to the designated charging facility.
Power System Model
To implement a predictive charging strategy for electric vehicles (EVs) within a VANET-enhanced smart grid, the power flow on distribution feeders is explicitly considered This work presents a power-system model that expresses the relationship between bus voltages and power loads, providing the foundation to derive how EV charging loads on feeders interact and affect feeder-level constraints By linking voltage profiles to EV charging demand, the model enables accurate forecasting of charging loads, informed scheduling, and mitigation of voltage deviations and feeder congestion The result is a coherent framework for coordinating EV charging in real-time grid operation, improving grid reliability and efficiency in VANET-enabled smart grid environments.
Consider a smart grid whose system model is shown in Fig 3.1, where the power system is abstracted as a one-line diagram with multiple buses For illustration, an example of a 12-bus system is presented in Fig 3.2a, and Fig 3.2b shows the abstracted equivalent power system model of Fig 3.2a Denote the set of buses in the system as N, with |N| = 12 in this example The generation buses are defined as the buses injecting power into the network.
1 DSRC protocol supports both RSU-to-vehicle/vehicle-to-RSU (R2V/V2R) and vehicle-to-vehicle(V2V) communication.
Fig 3.2 The power system model (a) Illustrated power system model (b) Equivalent power system model
Figure 3.3 illustrates the power flow of the system, with Bus 1 (shown in Figure 3.2a) serving as the generation point and all other buses that carry only loads designated as load buses (for example, Bus 3 and Bus 6) The power system is supplied via the substation connected to the generation bus EV charging stations are co-located with the load buses, such as at Bus 3 and Bus 6.
Bus 3 ,Bus 6 ,Bus 9 andBus 12 , respectively Each charging station is connected to the grid via a standard single-phase Alternating-Current (AC) connection Due to the thermal limit of service cable or current rating of fuse, an EV charging station at
Bus j is constrained by its total load capacity, P_j,total [9] Although a local vehicle-to-grid concept exists [10], bidirectional power flow and the EV battery’s directional discharge are not considered in this chapter The voltages of two adjacent buses during period k, for example V_i,k and V_j,k as shown in Fig 3.3, can be approximated according to the model described in [29].
During period k, the voltages at buses i and j are Vi,k and Vj,k, while Pij,k and Qij,k denote the active and reactive power flows from bus i to bus j in that period The impedance of the feeder line between i and j is rij + j xij Equation (3.1) expresses the relationship among these quantities, linking the voltages, the line impedance, and the power flows When expressed in per unit, (3.1) can be approximated in a simplified form suitable for practical analysis.
V i ; k V j ; k DP ij ; k r ij CQ ij ; k x ij : (3.2)
All bus voltages must remain within prescribed bounds, which constitute the primary operational constraint of the distribution system [29] For example, the voltage magnitude at Bus j during period k is bounded by an upper limit V_j^{max,k} and a lower limit V_j^{min,k}.
28 3 Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid
On each line k, the voltages V_j are bounded by V_j^min;k and V_j^max;k, i.e., V_j^min;k ≤ V_j;k ≤ V_j^max;k It has been proven in [29] that the minimum voltage point can occur only at the end of the power line, since the only generation bus is located at the beginning of the distribution system Therefore, the minimum voltage V_N;k can be derived as
V N ; k DV 1; k N P 1 iD 1ŒP i iC 1/; k r i iC 1/CQ i iC 1/; k x i iC 1/: (3.3)
EV Mobility and Charging Model
Two random variables, V and D, are used to characterize the mobility of each EV V represents vehicle velocity and can take two values: a lower velocity vL and a higher velocity vH The velocity transitions are modeled as a two-state continuous-time Markov process with a state-transition rate D1 In this model, a vehicle initially adopts vL (or vH) and, after an exponentially distributed waiting time with mean D, switches to vH (or vL), respectively.
The model captures realistic human driving behavior by allowing drivers to maintain a constant velocity for a period and then switch to another speed based on personal intent or road conditions It also shows that at low to medium vehicle density—no more than 30 vehicles per kilometer per lane—vehicles tend to move independently, indicating negligible interactions under these conditions Furthermore, the headway distance between vehicles follows an exponential distribution with a rate parameter, reflecting the stochastic spacing observed in these traffic scenarios.
When a mobile EVv(2 V) is charged atBus j in period k, the charging load
(denoted asPch v; j ; k ) should be within a certain range to protect the EV battery and power system stability, i.e.,
0Pch v; j ; k Pch max v; j ; k (3.4) wherePch max v; j ; k is the pre-fixed charging load upper bound ofPch v; j ; k [34] If EVvis not scheduled to be charged in periodk, i.e.,x v; j ; k D0,Pch v; j ; k should be0, i.e.,
2 Note that if the distributed generation is adopted in the distribution system, the overloading problem should also be considered.
3 In this chapter, the headway distance is defined as the distance between two neighboring vehicles in the same lane.
(3.6) whereX max is the maximum total charging times for an EV within all the considered periods, since frequent charging is not desirable and may cause battery damages
During a charging period, the energy supplied to each electric vehicle should be capped by its battery’s maximum capacity (Cmax_battery), and the battery should not be depleted on the way to ensure that the EV can be charged successfully.
Let P_init(v,k) denote the energy stored in EV v at the beginning of period k, which can be obtained via VANETs, and let P_cost(v,k) denote the travel cost for EV v to charge in period k The analysis is constrained by the battery capacity C_max, the maximum energy the vehicle’s battery can hold (as in equation 3.7) Notation P_k^cons denotes the average non-charging energy cost for each EV in period k.
EV while moving on the road if the EV is not scheduled to charge in period k.
The time horizon is divided into equal periods, each with duration τ hours For example, if each period lasts 0.5 hours (30 minutes), then Δt = 0.5 At an EV charging station located at bus j, the total EV charging load P_ch^{j,k} in period k is defined as the sum of the charging demands of all connected vehicles during that period, constrained by the station’s capacity and any scheduling rules This time discretization enables precise tracking of charging demand across the network and supports period-by-period optimization of charging schedules.
Transmission Model in VANETs
To enable V2V and V2R transmissions in VANETs, the draft IEEE 802.11p standard (DSRC) is used, a protocol designed for short-range, intermittent vehicular communications among vehicles and roadside units For analytical simplicity, an ideal MAC is assumed, eliminating interference among V2V transmissions; as soon as a vehicle enters an RSU’s coverage, the RSU can schedule nonoverlapping time slots for V2R exchanges without collisions We also assume a constant transmission rate for both V2V and V2R links and that each contact duration is long enough to deliver a single packet, achievable by selecting an appropriate packet size Given high mobility and intermittent connectivity, the waiting time to seize a transmission opportunity becomes the dominant component of transmission delay, exceeding queueing delay and the random backoff time from channel contention, so the analysis focuses on this primary delay term.
30 3 Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid
Problem Formulation
Charging Load Constraints
The charging station atBus j has a load-capacity constraintP j total , i.e., the total EV charging load atBus j in periodkshould be no more thanP j total Thus we have
Bus voltage declines as the load increases, and when a bus voltage drops below a threshold, reactive power cannot be injected efficiently To keep voltages within a target range, the load must be kept under a desired level, revealing a fundamental trade-off between voltage profiles and load allocation In this setting, Theorem 3.1 formalizes the essential relation among the EV charging load capacities across buses, and its proof is derived from power-flow analysis of the power system.
Theorem 3.1 (Linear relation among EV charging loads of buses) Given the total supplied power from the feeder and the non-EV charging load, the total power supply for all EV charging stations can be obtained The power supplied for one individual charging station has alinear relation with that of the other charging stations.
Proof For each bus, the voltage should be no less than the minimal required voltage, e.g.,0:9per unit voltage [9] According to (3.3), the lowest voltage isV N ; k ofBus N Then we have,
V N ; k DV 1; k N P 1 iD 1ŒP i iC 1/; k r i iC 1/CQ i iC 1/; k x i iC 1/ V min (3.10)
3.3 Problem Formulation 31 whereV min is the minimal required voltage Re-arranging Eq (3.10), we can get
N P1 iD 1ŒP i iC 1/; k r i iC 1/CQ i iC 1/; k x i iC 1/ V 1; k V min : (3.11)
Let w denote the sorted index of the bus with no EV charging load, and let W be the set of these buses, with w ∈ W ⊆ {1, , N} Denote the sorted index of the bus with EV charging load by j, and let the corresponding set of these buses be H, with j ∈ H ⊆ {1, , N} Then, (3.11) can be expressed by partitioning the bus indices into the non-charging group W and the charging group H, according to their sorted indices w and j.
Let P_j,k and Q_j,k denote the active and reactive power loads on bus j during period k, respectively The feeder impedance from Bus 1 to Bus j is described by its average resistance r_j and average reactance x_j, which capture the line’s loss and voltage-drop characteristics along the feeder These parameters are used to model how power flows and voltages respond across the network Since the forecasts provide the loads on buses without EV charging, we have a known baseline for those buses, enabling assessment of the incremental impact of EV charging on the rest of the system.
As each charging station is connected to the grid through a single-phase AC connection and EV charging only draws active power, we haveP j ; k D Pch j ; k and
Q j ; k D0 Thus, we can re-write (3.13) as
Equation (3.14) introduces a constant r_j as the right-hand side of inequality (3.13) This inequality shows that the locations and the total number of EV charging stations play a crucial role in the total power supply available to charging stations Given the total power supply to EV charging stations, the total load at bus j in period k, P_ch_j;k, presents a linear relationship with the other variables.
Travel Cost for EV Charging
Section 3.2.3 defines the charging-period travel cost for each electric vehicle v at period k, denoted as C_v^k, as consisting of two parts The first part captures the travel cost arising from the distance between the vehicle’s current location and the chosen charging station in period k, represented by d_v^k Since mobile EVs may occupy different locations and have varying battery levels across periods, this distance—and thus the travel cost—varies over time To alleviate range anxiety, drivers tend to select nearby charging stations with shorter travel distances, making the travel distance a major determinant of the EV charging travel cost.
32 3 Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid
The second component of travel cost arises from the transmission delay incurred when an electric vehicle receives a charging decision via VANETs, denoted as a_s_vk Vehicle mobility induces intermittent transmission delays in VANETs, causing fluctuations in V2V and R2V connectivity These intermittent links can introduce additional transmission delay and extend the travel distance the EV must cover before obtaining the charging decision from roadside units (RSUs) or nearby vehicles Therefore, transmission delay must be factored into the evaluation of travel cost.
1 Travel cost due to the EV travel distance to a charging station: If EV v is scheduled to be charged in the next periodk, thenP j2H x v; j ; k D1 The shortest path algorithm [40] is exploited to calculate the traveling path for EVvto the charging stationjin periodkby the installed GPS Denote the path length asS.x v; j ; k / Thus, the travel distance of EVvin periodkfor charging can be expressed as p v; k D P j2H S.x v; j ; k /x v; j ; k : (3.15)
Given p v; k , define the travel cost in terms of energy of EVv in period k as
PC.p v; k / The notationPC./is a linear non-decreasing function that measures the impacts of travel distance on the travel cost [10].
2 Travel cost due to the transmission delay in VANETs: The second part of the travel cost is introduced by the transmission delay of an EV to send (or receive) the charging request (or decision) to (or from) the neighboring vehicle or nearest RSU.
To quantify the transmission delay of the last hop in the V2R path, we attribute the delay primarily to the inter-contact time between the vehicle and the RSU We model the last-hop transmission with an on–off framework: the vehicle connects to the RSU during the “on” state, or, as the first vehicle approaching the RSU with no other vehicles within the RSU’s transmission range, it is in the “off” state Because the transmission delay in the “on” state is far smaller than in the “off” state, the overall delay is dominated by the “off” period Both the “on” and “off” periods are random variables, denoted as T_on and T_off respectively.
Let U_on and U_off denote the travel distances within the on and off periods, respectively, consistent with the mobility model’s average velocity V as defined in Section 3.2.3 Inspired by [31], the event that a vehicle moves a distance of at least u during T_on before being scheduled to transmit with an RSU satisfies two conditions: (1) no other vehicles are within distance u from the considered vehicle, and (2) the transmission opportunity is available.
3.3 Problem Formulation 33 another vehicle is within the distance2Ruso that the considered vehicle has to move at leastudistance to avoid the collision Here,Ris the transmission range of an RSU or a vehicle Then, the probability thatU on >uis calculated as
Let R denote the total length of roads and b the vehicle density on those roads In equation (3.16), these two quantities determine the relationship of interest As noted in Sect 3.2.3, the vehicle headway distance H follows an exponential distribution, so the probability that a headway exceeds u is P(H > u) = exp(-b u) Consequently, the mean headway, μ = E[H], is 1/b, providing the average spacing between vehicles on the road network.
During the off period, a vehicle can cover a distance of at least u only if two conditions are met: first, there are no other vehicles within a small clearance distance from the end of the nearest RSU’s coverage ahead of it; and second, there is at least one other vehicle within the inter-RSU spacing, L, where L is the distance between consecutive RSUs Under these conditions, the average travel distance during the off period can be determined.
“off” period can be calculated similarly as
In the preceding hops of the transmission path, the links are vehicle-to-vehicle (V2V) connections that can be described by a vehicle mobility model The evolution of the relative velocity between two adjacent vehicles is modeled as a continuous-time Markov chain (CTMC), with a state space that captures the range of velocity differences and their transition dynamics.
Let h0, h1, and h2 denote the three relative-velocity states between consecutive vehicles: h0 for a negative relative velocity when the vehicle ahead moves with vL and the vehicle behind with vH; h1 for zero relative velocity when both vehicles travel at the same speed; and h2 for a positive relative velocity If the duration that each vehicle maintains a given velocity is exponentially distributed with mean D, the CTMC transition rate between any two states is 2/D Accordingly, and as shown in [31], the average number of hops M along the entire message transmission path can be estimated from this Markovian model.
Based on Eq (3.20), the transmission delay of the whole transmission path can be given as
34 3 Mobility-Aware Coordinated EV Charging in VANET-Enhanced Smart Grid where EŒT V 2 V is the average transmission delay for a V2V hop satisfying that
EŒT V 2 V D 1 e 1 R , since the headway distance follows an exponential distribution.
Considering request sending and decision receiving as a unified process, the total transmission delay from issuing a charging request to receiving the charging decision is two time units This delay depends on network parameters such as vehicle mobility (low-speed v_L, high-speed v_H, and distance parameter D), vehicle density, and RSU deployment in the network (the transmission range R and the average inter-RSU distance L).
In period k, the average travel distance for EV v, while both moving and waiting for the charging decision, is computed as c_v,k = Ω_v ∑_{j∈H} x_{v;j,k}, where Ω_v denotes the travel distance attributable to the transmission delay of VANETs The corresponding travel energy cost is defined by a linear, non-decreasing function P_C(c_v,k), which measures the energy expended by EV v while waiting for the charging decision in period k.
Let P_C^p(v_k) and P_C^c(v_k) be the defined energy costs for the p and c segments at state v_k The initial stored energy P_init(v_k) must be at least the sum of P_C^p(v_k) and P_C^c(v_k) to guarantee the vehicle can reach the destination charging station.
P v; cost k DPC.p v; k /CPC.c v; k /P init v; k : (3.23)Note thatP init v; k can be collected in a real-time way by RSUs via VANETs.
Mobility-Aware EV Charging Optimization Problem
By exploiting a linear relationship between the load capacities of charging stations and EV travel costs, the charging strategy is designed to maximize the overall charged energy minus travel cost while preventing power-system overload The objective function encodes a joint priority: grow the total charged energy while minimizing EV travel costs, with a deliberate balance between the two Specifically, upon receiving (1) historical readings from RTUs installed at the buses and (2) vehicle information via VANETs, the traffic server determines the charging decisions P_ch^{v}_{j,k} and the decision variables x^{v}_{j,k} by solving the optimization problem presented below.