A Smart Approach to Electric Vehicle Optimization via IoT-Enabled Recommender Systems

Abstract

Electric vehicles (EVs) are becoming of significant interest owing to their environmental benefits; however, energy efficiency concerns remain unsolved and require more investigation. A major issue is a lack of EV charging infrastructure, which can lead to operational difficulties. Effective infrastructure development, including well-placed charging stations (CS), is critical to enhancing connectivity. To overcome this, consumers want real-time data on charging station availability, neighboring station locations, and access times. This work leverages the Distance Vector Multicast Routing Protocol (DVMRP) to enhance the information collection process for charging stations through the Internet of Things (IoT). The evolving IoT paradigm enables the use of sensors and data transfer to give real-time information. Strategic sensor placement helps forecast server access to neighboring stations, optimize vehicle scheduling, and estimate wait times. A recommender system is designed to identify stations with more rapidly charging rates, along with uniform pricing. In addition, the routing protocol has a privacy protection strategy to prevent unauthorized access and safeguard EV data during exchanges between charging stations and user locations. The system is simulated with MATLAB 2020a, and the data are controlled and secured in the cloud. The predicted algorithm’s performance is evaluated using several kinds of standards, including power costs, vehicle counts, charging costs, energy consumption, and optimization values.

1. Introduction

Recently, countries worldwide have enacted regulatory measures to tackle challenges within the transportation industry, including the growing demand for fossil fuels and urban pollution concerns. Eco-friendly EVs have drawn a lot of interest due to worries about the availability of fossil fuels, the environment, and emission restrictions. Numerous nations have set forward significant rules to facilitate the growth and adoption of EVs [1]. For instance, some cities in North America are undergoing a rapid transition from conventional vehicles to hybrid, plug-in vehicles and, subsequently, to EVs. EVs provide many advantages in terms of price and convenience. An electric vehicle (EV) does not require any maintenance, such as the expense of lubrication, the exhaust system, starter battery, or cooling fluids, as opposed to an internal combustion engine (ICE) car [2]. EVs’ driving range is also expanding more quickly than in the past. With a 16-kWh rated capacity, the small Mahindra Reva car can travel 80 km on a single charge. Tesla Model X (SUV) vehicles with a 94-kWh battery pack have a range of 400–450 km.

Figure 1 depicts a typical electric vehicle charging station and Figure 2 shows a block diagram of an electric vehicle charging station that includes a rectifier for converting AC to DC, a DC/DC converter is connected to boost the voltage according to the charging station’s capacity to enable the effective charging of the EV’s. For further safety, the system also incorporates a protection control mechanism in the vehicle and charge controller in the charging station. The size of their batteries primarily governs EVs’ driving ranges, but engine complexity also plays an important role. Because recent research has concentrated on extending the driving range of EVs, battery monitoring and management technologies are necessary. These offer a thorough rundown of the Li-ion batteries found in electric vehicles. For high-capacity batteries, which call for a quick charging system, the fundamental operation ideas, assembly, and performance of several types of Li batteries are addressed and recommended. Rapid CSs can charge a battery from a 20% state of charge (SOC) to approximately 80% in 30 min, according to [3]. On the other hand, level II and level I require an extended charging process. A public CS is a preferable option for charging for EV owners who do not have access to their chargers. EVs and hybrid electric vehicles (HEVs) [4] have assessed the ideal charger mode, ensuring efficiency and the quickest charging time.

SOC is one of the crucial factors that makes it possible for the battery to be charged appropriately and discharged, extending battery life. This word refers to the battery’s rated capacity ratio and balancing capabilities. SOC, hence, aids in battery management. Various SOC estimating methods are provided in [5]. Three categories exist for new communication systems in the automobile industry: vehicle-to-driver (V2D), vehicle-to-vehicle (V2V), ref. [6] and vehicle-to-infrastructure (V2I). Governments have made significant investments, so communication technologies like V2V and V2I have undergone extensive research. These communication technologies are primarily intended to lessen traffic congestion, increase road travel safety, and prevent vehicle collisions. Still, only some studies support vehicle-to-device systems for EVs [7]; further studies in this field are required. Intentionally assisting the EV user and CS with effective power management scheduling is data communication. The closest CS receives the estimated SOC information and allocates the slot to the grid’s power demand [8].

EV load forecasting is crucial for utility companies to run and manage effectively. According to [9], the charging profiles are predicated on daily vehicle charging beginning at 17:00, 18:00, or 19:00 h. Actual EV charging loads, however, will be significantly influenced by travel routines, which differ considerably from driver to driver and from day to day. Dong et al. [10] predict the EV charging patterns for various locales using GPS data from 76 automobiles. However, the study assumed that all EVs were sedan-style vehicles with comparable features. The fleets’ plug-in comprises 50% mild hybrid EV charging loading, 20% fully EVs, and 30% fully hybrid for EV charging loads [11]. According to the parking period, Ref. [12] simulated the demand for EV charging by considering the spatial and temporal distribution. The author describes the investigation and performance of EV load forecasting by analyzing historical traffic and weather data and the classification of EV charging commission using the decision tree algorithm. However, this analysis only considers vehicles with 27 kWh Li-ion batteries, and it was also believed that a car could only be charged once daily.

Given that a substantial portion of decisions in the power and energy domains hinge on prognostications of future requirements, the authors employed various methodologies, encompassing the Monte Carlo forecasting approach and support vector machines, to gauge the anticipated demand for EV charging [13]. A reliable evaluation metric for electrical load forecasts in V2G scheduling considers statistical features of EV charging. Multiple techniques were utilized for short-term prediction, including the akin-day method, time series analysis, regression models, expert systems, neural networks, statistical learning and fuzzy logic algorithms. The advancement, enhancement, and scholarly exploration of pertinent mathematical instruments can facilitate the formulation of more precise load forecasting techniques. Based on an accurate real-time range indicating system and traction power assessment, Sadiq et al. [14] offer an EV driver a real-time charging recommendation. However, with the current static system, the driver cannot learn whether the CS is operational, how many vehicles are awaiting charge, or which charging points are available. The works mentioned above revealed the need for a real-time application to gather transportation data, EV battery condition, and CS status to estimate potential loads on CSs and avoid relying on aggregators [15]. This research introduces a server-based real-time forecasting application aimed at achieving two primary objectives: (i) optimizing scheduling to minimize wait times and (ii) delivering real-time charging station recommendations for EVs that combine affordability and rapid charging [16,17]. The following are the main contributions of this suggested initiative:

(1) The developed application for EVs needs the closest CSs with names, positions, and geographical route coordinates and reserves the slot to charge the EV before arriving at the destination. The algorithm does not need any complex data interchange. For the driver to find the CS, everything can be automatically retrieved via the Distance Vehicle Multicast Routing Protocol (DVMRP) and the internet system via openly accessible websites.

(2) The suggested application is intended to be sufficiently intelligent to provide the driver with appropriate knowledge about the CSs while lowering the risk of station location with other vehicles, all with minimal data input from the driver.

(3) This work proposes a flexible, real-time-based online charging scheduling scheme where each moving EV can choose, instead of depending on the selection from the aggregator. This flexible real-time-based online charging scheduling scheme considers the available efficiency, packet delivery ratio, delay and throughput.

According to this perspective, our suggested system is desirable because the user solely makes the decision, and the privacy of EVs will not be revealed through any channel. In addition, the scheme offers information about the accessibility of slots at the closest CSs, as well as cost and distance based on battery parameters. According to the vehicle’s SOC and the vendor’s viewpoint, the management of the energy system can respond based on the CS status, which is communicated through websites. The webpage is created using a hypertext preprocessor so that the user can view the necessary information. Therefore, users can choose CS based on distance and cost per unit. The database is created using cloud SQL. The relevant data, like the price, SOC, average rate, and so forth, will be in the database. This database is linked to the webpage and uses the internet to provide the user with the information. A real-time application that gathers transportation data, EV battery state, and CS status (cost/kWh, range, total slots, and availability of fast charging option) is constructed to anticipate potential loads on CSs accurately. In conclusion, today’s stakeholders, such as technology users and developers, who would expect this facility in their vehicles, can be very interested in introducing an intelligent system between cars and drivers. It is anticipated that these technologies will boost the uptake of EVs in the automotive sector, especially in developing nations.

The rest of the paper is organized in the same way. The current work is elaborated in Section 2. Section 3 discusses the EV charging infrastructure and the intended IoT application. The outcomes and the IoT-based applications are covered in Section 4. Section 5 contains conclusions from this work, a summary of the work performed, and work to be performed in the future.

2. Literature Review

Lei et al. [18] have developed a mobile Android application that conveniently displays the nearest charging stations (CSs) in proximity to electric vehicles (EVs). To facilitate slot reservations, the application employs intelligent features to recognize real-time CS status, including slot availability and pricing, while also providing a list of available CSs and an estimated driving range. Similarly, the author investigated State of Charge (SOC) estimation and CS recommendations. Their proposed technique alerts EV drivers when the safe SOC limit is reached and provides accurate estimates of charging duration, enhancing the overall charging experience.

Nakakohara et al. [19] contribute to the field with another mobile Android application that displays CS locations physically closest to the EV. While offering a list of available CSs and a driving range estimate, this application requires intelligence to dynamically assess real-time CS status, such as slot availability and pricing, for effective slot booking. Xue L et al. [20] provide valuable insights into designing and implementing a Dual Active Bridge-based battery charger tailored for PHEVs, focusing on mitigating low-frequency ripple in the charging current. Its findings contribute to advancing charging infrastructure for electric vehicles, addressing critical challenges in efficiency, reliability, and performance.

To improve the performance of DAB converters with phase-shift control, Rodríguez Alonso et al. [21] offer insightful information that advances power electronics technology in various industrial and automotive applications. Huang et al. [22] emphasize the importance of automated communication infrastructure, such as the internet or wireless networks, for a charge controller to receive EV user information. This information allows the charge controller to manage EV charging based on real-time data. Every instance of EV plug-in or out, as suggested by [23], necessitates notifying the aggregator of available SOC and anticipated departure time for optimal charging management. Ref. [24] achieves flexible EV charging by exchanging information with three distinct operators: charging service provider (CSP), retailer (RET), and distribution system operator (DSO). In Ref. [25], hardware implementation of a plug-in EV wireless charging manager is described as connecting the three stakeholders. However, a detailed description of the communication system between EVs, the store, and the DSO is needed.

The paper [26] presents a novel strategy to schedule the charging points. The goal is to determine the convenient CS for a given EV through a Vehicular Ad-hoc Network (VANET) model. In this model, the CSs are determined and prioritized using four phases, such as driving, charge planning, charging scheduling, and battery charging. Charging scheduling was designed using a new optimization strategy, exponential Harris Hawks optimization (Exponential HHO) algorithm. The proposed Exponential HHO was validated using simulation and the performance was improved with maximum remaining energy of 52.709 Whr, minimal distance of 27.256 km, and a maximum average waiting time of 0.352 min in comparison with existing methods.

The authors of [27] provide a comprehensive approach to forecast the mid-and long-term charging load of electric vehicles (EVs), with a focus on the Chinese city of Shenzhen. The methodology integrates several strategies to precisely forecast future electric vehicle charging requirements. Gathering data on EV usage patterns, including travel distances, charging routines, and parking durations, is the initial step in the process, which involves intensive data collection and preprocessing. Then, preprocessing is carried out on the information in order to ensure consistency and standardization. The author has developed a systematic methodology to forecast the future EV charging loads for China’s next five-year plan (2020–2025). For better evaluation, the entire EV fleet is divided into four categories (namely, private EV, electric taxi, electric bus and official EV) and the charging profiles of EVs are established according to real-world data and Monte Carlo simulations are employed to obtain the future load demands resulting from EV charging [28,29].

To address the computational challenges in EV load forecasting, an aggregate model of the EV fleet is introduced in [30]. This model produces a representative scheduling result by utilizing the statistical features, without heavy computation. Subsequently, a reliable metric from the perspective of forecast value, called V2G scheduling value error (V2G-SVE), is proposed. Empirical case studies affirm the reliability of V2G-SVE as an evaluation metric for load forecasting. Also, this paper reveals three valuable findings regarding the relationship between forecast errors and V2G scheduling performance, accompanied by an illustrative mathematical proof, illuminating potential enhancements for future load forecasting technology.

In [31], a price-responsive billing structure is used for effective online scheduling. Zhang et al. [26] address mobility uncertainty through bidirectional communication, allowing an EV aggregator to manage charging controls for an EV fleet. Information and communications technology (ICT) gathers information on the EV fleet for optimal charging in the energy market. Ref. [32] utilizes modern communication technology to transmit EV information to the energy provider for synchronization of charging and discharging.

However, these systems present potential security issues, especially in protecting private information like vehicle registration numbers [33]. EVs frequently seek decisions from global aggregators [34], potentially exposing information to malicious entities. Privacy concerns, such as targeting EVs with unsolicited items or services based on location, underscore the importance of safeguarding EV charging status [35]. EV users must be well-informed about the availability of the closest CS to avoid making incorrect choices [36].

3. Infra-Structure Analysis

3.1. Topology Setting and Charging Station Concepts

The distribution network can comprise energy storage, local renewable energy sources (RES), and several chargers. It can also run in either an AC or DC configuration. The ensuing sections delineate the obstacles and prospects and the methods employed for deploying certain kinds of charging station. Various power electronics converters are identified and compared for use in these applications. This work discusses both the advantages and disadvantages of these approaches. Additionally, this work explores the suggested topology modifications and control enhancements from the existing literature to better align with Extreme Charging (EC) requirements. The converter topologies appropriate for EC applications are the main subject of this work; onboard charger topologies still need to be covered [37].

3.2. AC and DC Power Distribution

Step-down transformers act as a conduit between the distribution network and three-phase AC buses in AC-connected systems. These buses operate between 250 and 480 volts line-to-line. Each charger at the station is powered by an AC bus and features its own independent AC/DC stage. The number of conversion stages between the DC port of a RES or an EV, like a battery or photovoltaic system, and the distribution network is significantly increased by this method [38]. In AC-connected systems, a higher number of conversion stages can lead to reduced system efficiency, increased system costs, and heightened system complexity [39]. Adopting the AC bus has numerous advantages, such as better and easier-to-access rectifier and inverter technologies, readily available switchgear and protective devices, established standards and protocols for AC power distribution systems, and much more.

Figure 3 shows the block diagram of the AC charger where, AC to DC conversion happens in electric vehicles. Figure 4 shows a block diagram of a DC charging unit, where the conversion happens at the charger in which the DC-connected devices with energy storage and RESs can operate more energy effectively since a DC bus can be built with just one central front-end AC/DC converter. A Solid-State Transformer (SST), sometimes called an LV (250 V–480 V) rectifier stage, comes after the low-frequency transformer at the front end. SSTs integrate isolation, voltage step-down, and rectification processes into a single device. To be compatible with the current battery voltage range of about 400 V, the DC bus voltage typically stays below 1000 V.

The design requirements for EC stations with DC and AC buses should be the same at this voltage level. Connecting each charger to the DC bus via a DC/DC converter reduces the number of conversion steps compared to AC-connected systems [40]. It eliminates the need for separate AC/DC converters. The “DC distribution” approach may have the benefit of only requiring a single central front-end connection to the utility. Additionally, by using this technique, the nameplate rating of the grid connection and the AC/DC converter can be significantly lowered, which will ultimately result in a reduction in the overall cost of system installation. It is accomplished by taking advantage of the load diversification generated by varying EV battery capacity and SOC-adjusted variances in battery charge acceptance.

Furthermore, another advantage of DC systems is their reduced need for control, as they do not involve reactive power considerations. It is simpler to connect to an island system from the primary grid because of the single inverter’s direct grid connection. Through partial power converters, DC distribution systems may make communication between the car and the DC bus easier. These partial power converters only process a portion of the power supplied to the vehicle, which lowers converter ratings and boosts conversion efficiency [41]. For instance, using several partial power DC/DC converters to communicate with the common DC bus of an EC station is recommended [25]. These converters prevent galvanic isolation between cars by sending a portion of their control power straight from the DC bus to the vehicle. Since “each output circuit must be completely isolated from each other” for an EV charging station, this technique must overcome several significant technological challenges to comply with current charging rules.

DC-connected systems offer certain advantages but have unique challenges, including DC metering and protection issues. Various safety devices are available for low-voltage DC systems, including protective relays, solid-state circuit breakers, fuses, and circuit breakers. There are not any set criteria for protective coordination for DC-connected EV charging stations yet. Numerous complex factors, such as fault type, system topology, grounding configuration, component specifications, size, and more, affect the protection coordination in DC-connected systems. When working with bidirectional chargers, the problem gets much more complicated. With minimal inertia, a DC-connected system is susceptible to disruptions and may become unstable if errors are not promptly fixed.

Thus, a critical factor in system restoration is the speed at which problems are found and isolated. Studying current DC power distribution systems, including low-voltage DC microgrids, can help one understand how DC-connected charging stations are protected from one another. For example, the model describes a low-voltage DC microgrid protection plan coordinating several preventive measures. An alternative method for safeguarding DC systems using a loop-style bus is described. Isolated faults can be recognized, and the proposed plan keeps the power supply intact. In a DC-connected system, it is essential to install DC meters to measure energy generation and consumption associated with EV chargers, battery energy storage, and RES [10]. In addition to being useful for future station development, these data are essential for the appropriate billing of EV station consumers. Despite being readily available, DC meters need standardized accuracy, calibration, and testing protocols that permit their use in metering applications. For systems that are connected to DC, the development of such standardized and approved DC meters is required.

3.3. Bidirectional AC/DC Converters

Figure 5 shows the bidirectional AC/DC converter in which the supercapacitor enables rapid energy release, hence fast acceleration, while the battery provides higher storage and, therefore, a long range. The most popular option for the grid-facing AC/DC converter is a three-phase active pulse-width-modulated (PWM) converter with an LCL filter. The input’s line-to-line peak voltage is less than this boost-type converter’s output voltage. In addition to enabling bidirectional power flow and generating low harmonic input currents, the six-switch PWM converter allows for modifying any power factor (PF). The most cutting-edge DC fast chargers often adopt this architecture due to the simplicity of its layout, ease of management, and accessibility to reasonably priced IGBT devices with adequate voltage and current ratings. The neutral point-clamped (NPC) converter displayed is an additional illustration of a boost-type converter implementation.

By employing lower voltage-rated components, the three-level converter can effectively reduce switching losses cost-effectively. In addition, it generates a three-level voltage waveform that lowers dv/dt and input current harmonics. A prototype 30 kW EV charger with an NPC front-end creates minimal input current total harmonic distortion (THD) by employing the input transformer’s leaking inductance as an AC-side filter [13]. Explicitly building a bipolar DC bus with the NPC converter as the AC/DC front-end has an extra benefit. A bipolar DC bus can create an EV charging station because the DC/DC converters can connect to half of the DC bus voltage. These two papers go into great detail about this functionality. A bipolar DC bus is another way to incorporate partial-power converters into the DC/DC stage [26].

3.4. Charging Application

SSTs frequently use similar modules as building blocks to attain the necessary voltage and power levels. The modules are linked in parallel at the output to supply a substantial output current at the required low DC voltage for direct connection with the medium-voltage (MV) grid and in series at the input to improve voltage-blocking capabilities. In this manner, the MV fast charger design was employed. Three modules are connected in parallel on the battery side and series (2.4 kV) on the MV AC side. For power factor correction and AC/DC conversion, each module has a unidirectional NPC AC/DC front end. Each module includes an internal bus at 1250 V to use SiC MOSFETs or commercial silicon IGBTs. Following the unidirectional NPC AC/DC front end, two input-series-output parallel phase-shift full-bridges extract the desired output voltage from the internal DC bus voltage [33,34,35]. Because of its heavy reliance on active switches, this design is less likely to achieve good compactness and efficiency and increases system costs. It is one of the design’s drawbacks. The phase-shift complete bridge also provides a fixed 450 V output. An additional DC/DC stage is required if the converter is connected to an electric vehicle to meet the battery charging profile. From the perspective of the battery, the electric vehicle battery is integrated into the architecture through a six-phase interleaved boost converter. This 38-kW system has an efficiency of over 96%.

3.5. Routing

Using distance vector information, the proposed upgraded DVMRP (Distance Vector Multicast Routing Protocol) distributes data about participating EVs across the network, including their location, vehicle ID, distance and charge state. Comparably, the embedded sensor nodes of the participating EVs use this information to communicate with other EVs via hop count and BS, sharing the data they have collected during the network. Additionally, an administrator helps to monitor the web from a distance [34]. The proposed DVMRP (Distance Vector Multicast Routing Protocol) consists of a variety of components, including routing protocols, sensor nodes, prediction algorithms, and converters. This routing concept creates a network topological infrastructure in which all EVs share information. It allows drivers to process routing and pricing requests across the network. However, the existing routing mechanism transfers data from the source to the destination depending on hop count. While this technique allows for roaming depending on hop count selection, it does not provide effective communication inside the network infrastructure. In the traditional routing approach, communication of data from source to destination will depend on hop-count information. Although this permits roaming, it frequently fails to ensure efficient communication within the network infrastructure [37]. The modernized DVMRP for EVs aims to fix these issues by: (1) Using advanced routing techniques to optimize data routes; (2) Improving data prediction and processing techniques to better anticipate and control network traffic; and (3) To ensure safety and reliability, robust protection and charge control methods have been implemented. These enhancements are designed to achieve better network lifetime, minimizing delays in data transmission and enhancing overall performance and resource utilization.

To improve routing, the Distance Vector Multicast Routing Protocol (DVMRP) is used. Initially, DVMRP is set to zero, indicating that no pathways are created. The vehicle sequence is then reset to zero to prepare for the processing of electric vehicles (EV). The next step is to loop through each EV in the network, where n is the total number of EVs. Each EV, designated by a unique ID, is parked at a charging station M. For each EV, the protocol repeats the same steps. If an EV is not present, the DVMRP remains zero, indicating that there is no available path. If the EV is not at the charging station, it moves on to the next state in the sequence, and its count increases by one. Subsequently, communication is established with nearby charging stations to explore multiple routing paths. When the EV sequence is incremented by one, the sequence is updated based on the EV’s current state.

Since the remote administrator is crucial to maintaining the network’s dependability metrics, the administrator is essential in our suggested approach. Below is a more detailed explanation of how the enhanced DVMRP routing protocol is implemented in the proposed model—it first connects to adjacent EVs to update the routing table [35]. Here, EVs that are a member of the network link with other EVs through base stations (BSs) and communicate with close peers based on the hop count. Also, our improved DVMRP protocol will allow participating EVs to update their routing data when their positions change, while maintaining the same level of consistency in the parameters, like the proximity hop count data. Therefore, the EV guarantees continuous connectivity throughout the communication process, even when it roams to a different base station. It does this by maintaining accurate and constant communication within the network. The algorithm in the paper, Algorithm 1, comprehensively explains forwarding and storing data related to route replies (RREP) and route requests (RREQ) within the network.

Algorithm 1. Forwarding and storing data related to route replies (RREP) and route requests (RREQ) within the network.

Input: EV with minimal distance Output: Shortest distances of EV with ID Set $DVMRP\to 0$ Initiate $\to 0$ Initiate the vehicle sequence $\to 0$ For $i\to 0;i<n;i++$ Do For $i\to i\in EV$ with ID with charging station $M$ Do For $j\to$ successive sequence If $EV\to 0$ then $DVMRP\to 0$ //no path Elseif $EV\ne Charging station$ then return, $EV\to$ successive state;//$j+1$

$EV counts\to j+1$

Establish communication with other nearby stations; //multiple routes. End if End for If the sequence is $EV+1$ Then Update sequence based on EV

Let us consider a scenario where two EVs, A_i and B_i, belong to the group N_(n−1). EV Ai and EV Bi start a power exchange request in this network. After receiving RREQ from A_i, the B_i replies with an acceptable message to hand control. Note A_i will retransmit the RREQ packet to the network until B_i responds to A_i’s RREQ. Additionally, the remote administrator receives A_i’s RREQ. When B_i requests a power exchange across the network from AI, the distant administrator asks B_i to reply directly. Similarly, the remote administrator will request that B_i grant A_i’s request for power and assign a charge to A_i [36]. The EV (s) taking part in the enhanced DVMRP protocol can thus interact with each other and handle requests for power transfer within the network. The flow of the proposed model is based on the EV arrival rate, charging rate and routing performance. The cost and the availability of successive vehicles are also updated periodically, as shown in the flowchart in Figure 5.

3.6. Charge Estimation

Every participating EV uses the charge state estimator (CSE) to determine the amount of accomplished charging. The suggested scheme’s battery model was built on series connectivity, meaning all batteries have the same rating and charging capacity [29]. Similarly, wireless node embedding was further linked to the machine learning data that CSE gathered. The Enhanced DVMRP routing protocol connects EVs with embedded wireless nodes to the topological design of the network [19]. Drivers of these participating EVs can monitor local traffic conditions using the dashboard screens (DBC). Furthermore, if an EV is nearby, the driver can initiate and manage power transfer requests through the DBC interface. Power transfer involves the immediate sharing of network resources, including donor and acceptor EV (s), and CSE uses machine learning algorithms to capture the present situation. Consequently, each EV employs a generator (G_c) to generate magnetic fields (MF) for the activation of the magnetic coupling (MC). In particular, G_c makes it simpler to extract current from the donor EV and transfer charge by increasing current flow up to 50% from its initial value. This speeds up the charging process and lowers energy waste in the DWC configuration. Power banks from different EVs are charged to one another in this study.

3.7. Module Connection

The series-based battery connectivity is employed in our suggested method. However, each module’s (battery bank’s) charge is automatically balanced. When the magnetic connection between paired EVs is established, power transfer occurs. The EV then goes through a charging process to switch from a high SOC to a reduced SOC. Furthermore, given that the EV (A_i) battery bank is on the left, the EV (B_i) battery bank is located on the right of the figure. The B_i EV module is charging more smoothly, which is consistent with our assumption. B_i is asked to trade power by an artificial intelligence (AI) system, and B_i grants the request. Due to their series battery connection, B_i and AI modules will keep charging each other until their power levels are equal. Every battery module also has CSE that monitors the installed battery bank’s state of charging SC_l. The recommended approach’s series-based battery connectivity equalizes the charging condition. Additionally, a system that can entail connecting batteries in series to preserve each module’s maximum power level is used to equalize the modules M_(n−1), each of which has a state of charge SC_l (where “M” stands for a battery module).

Furthermore, the suggested model’s operation clarifies the idea of dynamic wireless charging or DWC. EVs approaching in parallel turn on their G_c to produce MF for MC, initiating the power transfer process. The power transfer process starts when the paired EVs establish an MC_c connection. Usually, this process ends with an EV in a low charge state (LC_l) and begins with an EV in a high state of charge (HC_l). Due to the G_c doubling property, which increases the initial electric current by 50%, the donor EV discharges the charge at a 50% higher rate. Fast and energy-efficient charge transfer between the paired EVs is made possible by the receiving EV’s acceptance of the discharged charge at the same accelerated pace. The low-power module M_l receives charge R_c from the high-power module M_h in one complete unit when M_h transfers control R_c to M_l. The charging state of the M_h module diminishes throughout the power transfer as R_c is released from M_h to M_l. Furthermore, the receiver EV’s M_l module’s charging status rises by (1−l_c)R_c. The battery cells B_i module charging system for EVs was selected for this study, with B as the battery’s i^th cell and S ∈ M as the module’s ith B. The cell n of B receiving a charge from its nearby cell B_i is represented. The charge transfer mechanism between the battery cells in the EV is stated in Equations (1) and (2). Power dissipation, which affects the way charge is transferred between cells, is represented by the term ‘smgl’. The cell state is taken into account by the equations, which show how charge is transferred depending on whether the cell index is within or outside of specified limits.

$$T_i^{cl}\left(n\right)=\left\{smgl \left(\left(-∆\right)\left(n-1\right), L_{m_n}\right)r_{m_n}, \left({\displaystyle \frac{i}{cell}}\right)-1>0\right\}$$

(1)

$$T_i^{cl}\left(n\right)=\left\{smgl \left(\left(-∆\right)\left(n-1\right), L_{m_n}\right)r_{m_n}, \left({\displaystyle \frac{i}{cell}}\right)+1\le \left(M\right)\right\}$$

(2)

Equation (3) calculates the difference in charge between the high charge state (H_c) and low charge state (L_c) over time, represented by ${∆}_{H_c}^M$. It helps in measuring the change in the charging state of the cells during power transfer.

$${\left(C_l\right)}_{H_c}^{\left(M\right)}\left(n-1\right)={∆}_{H_c}^M$$

(3)

Our scheme’s connection is displayed with an illustration of the generator and two connected coils to provide M_F for M_C. The transmitting coil charges the receiving ring after the connection is made. It is the method by which two EVs exchange electricity in our proposed method.

3.8. Mathematical Model

The integrated coil in every EV in the suggested model creates a magnetic field. As you switch from one EV to another for charging, the Gc will turn on. A magnetic field is created when the built-in coil for the DWC turns on. In a similar vein, the second EV that is sharing electricity turns on its G_c to create a magnetic field. In the suggested model, there is interference between the MF of two parallel EVs. The paired EVs transfer power to one another after the MC interference is established. Let us assume for the time being that there is the following mutual inductance between two coils:

$$Mutual inductance={\displaystyle \frac{N_1\ast N_2\ast {\mu }_0\ast {\mu }_r\ast A}{l}}$$

(4)

Equation (4) shows the calculation of the mutual inductance between two coils ($N_1 {\mathrm{and} N}_2$) where N₁ and N₂ are the number of turns in each coil, (l) is the coil length, (A) is the cross-sectional area and parameters μ₀ and μ_r are the permeability of the iron core and accessible space, respectively. Equation (5) determines the mutual inductance ${(M}_{12})$ between two coils using the magnetic flux, ${(\varnothing }_{21})$ through the second coil and its number of turns (N₂), where l₂ represents the length of the second coil. This equation indicates how well the magnetic field of one coil induces a voltage in the other coil.

$$M_{12}={\displaystyle \frac{{\varnothing }_{21}\ast N_2}{l_2}}$$

(5)

$$k={\displaystyle \frac{M}{\sqrt{coi{l}_1\ast coi{l}_2}}}$$

(6)

On the other hand, k in Equation (6) represents the coupling factor between two coils. The coupling factor measures the effectiveness with which the magnetic fields of the two coils are linked together. It is a dimensionless integer that ranges from 0 to 1. Similarly, an individual coil’s self-inductance can be found using the Equations (7) and (8), It helps determine how each coil stores energy in its magnetic field. To determine self-inductance and the mutual inductance between two coils, increase both in Equation (8). Equation (9) gives the mutual inductance M, in which the factor H is used to adjust the mutual inductance calculation based on the interaction between the coils.

$$coi{l}_1={\displaystyle \frac{N_1^2\ast {\mu }_0\ast {\mu }_2\ast A}{l}}$$

(7)

$$coi{l}_2={\displaystyle \frac{N_2^2\ast {\mu }_0\ast {\mu }_r\ast A}{l}}$$

(8)

$$M=H\sqrt{coi{l}_1\ast coi{l}_2}$$

(9)

3.9. IOT Framework for EV

The automotive sector has the potential to undergo a transformation owing to the integration of Internet of Things (IoT) technologies within Electric Vehicle (EV) systems. Real-time data processing, effective communication protocols, security measures, and scalability are some of the issues to be considered when designing an IoT (Internet of Things) framework for electric vehicles (EVs). When considering an IoT (Internet of Things) approach specifically for electric vehicles (EVs), the focus typically revolves around optimizing vehicle performance, enhancing user experience, ensuring safety, and enabling efficient fleet management. Performance Monitoring and Telemetry: Install sensors in the EV to gather data in real-time on a variety of characteristics, including speed, energy usage, environmental conditions, temperature, and health of the battery. IoT protocols can send this data to a centralized system for monitoring and analysis. Remote Service and Troubleshooting: Provide IoT connectivity to facilitate troubleshooting, predictive maintenance, and remote diagnostics for electric vehicle components. Potential problems can be found early on by continuously monitoring vehicle data, which minimizes maintenance costs and downtime. Infrastructure management charges considering features like automatic billing, intelligent charging scheduling, and remote charging station monitoring, integrate Internet of Things (IoT) capabilities into EV charging infrastructure. This optimizes charging sessions according to grid demand, electricity rates, and specific vehicle necessities.

Features for safety and security set IoT-based security measures to protect EVs from theft, illegal access, and cyber threats. This incorporates strong encryption and authentication methods along with features like geofencing, remote car locking and unlocking, and tamper detection. Implementing IoT solutions for driver behavior monitoring, fleet tracking, route optimization, and performance analytics for commercial electric vehicle fleets. Fleet managers benefit by increasing operational effectiveness, cutting expenses, and guaranteeing regulatory compliance.

4. Numerical Results and Discussion

Table 1. Parameter setup.

Types	Values
Environment	Simulation
Electric vehicles	10–50
$\mathrm{Generator} (G_c$)	1 per EV
Wireless nodes	Based on the total EV count.
Gateway	20 m
Base stations	5
Charge stations	Based on the total EV count.
Batteries	Based on modules
Routing protocol	DVMRP
Transmission range	1000 m
Charging environment	Charging state
PC	1
Packet size	128 kb

presents the parameter setup used for simulation. This section details electric vehicle hardware simulation outcomes with two, three, four, and six phases under distinct input voltages, precisely 20 V and 200 V. The focus is on efficiency and power loss computations at various stages, particularly the output voltage during the fall phases of the proposed model at a 20 V input.

Table 2. Output parameter value (i/p voltage 20 V).

No. of Phases	Voltage (V)	Current (A)	Power (W)
2	39	1.2	47
3	39	1.3	47.1
4	39.1	1.4	47.5
6	39	1.3	47.8

presents the Output parameter value such as voltage, current and power for the input voltage of 20 V) and

Table 3. Output parameter value (i/p voltage 200 V).

No. of Phases	Voltage (V)	Current (A)	Power (W)
2	435	13.8	5.9
3	436	13.5	6
4	437	14	6.2
6	438	14.8	6

presents the Output parameter value such as voltage, current and power for the input voltage of 20 V). In the case of the two-phase model, the output voltage stabilizes at 38.74 V within 0.2 s, as indicated by waveform data. A comparative analysis is presented for output current and power at a 20 V input, revealing a power output of 46.914 W and an output current of 1.211 A. Moving on to the three-, four-, and six-phase configurations, the respective output voltages settle at 38 V, 39 V, and 40 V within varying time intervals of 0.15, 0.11, and 0.05 s. Corresponding power and output current values for the third, fourth, and sixth phases are 47 W, 48 W, 47.5 W, and 1.215 A, 1.218 A, 1.22 A.

In addition to performance metrics, reliability measures are assessed using the anticipated model during experimental outcomes. This includes evaluations of charging efficiency, packet loss, end-to-end delay, and throughput. Notably, the packet loss ratio is employed to assess the effectiveness of the upgraded DVMRP protocol. In the communication process, electric vehicles utilize hop count and direct communication when a base station is involved. Live network performance reliability is verified by tracking the enhanced DVMRP protocol’s throughput and end-to-end delay statistics. These parameters are summarized in Table 1.

The output parameters derived from the MATLAB 2020a simulation at different phases are presented in Table 2 and Table 3. The results in Table 2 and Table 3 show that as the number of phases grows, so do the output characteristics—voltage, current, and power.

This observed pattern can be attributed to the diminishing switching stress resulting from the concurrent reduction in induction current. As the number of phases grows, the decline in induction current contributes to a reduction in switching stress. Consequently, the simulation results show a corresponding decrease in power losses and an increase in the efficiency of the models under consideration. This insight underscores the positive impact of a higher number of phases on the overall performance, suggesting improved efficiency and diminished power losses in the simulated models.

4.1. Efficiency and Power Analysis

Equation (10) gives the Input power equation $P_{input}$ in which P_output denote the output power, and power losses associated with diode, capacitor, and inductor are represented by P_diode, P_capacitor and P_inductor, respectively. From Table 2 and Table 3, it is observed the improvements in efficiency and reductions in power losses, highlight the positive impact of increasing the number of phases on overall performance. The efficiency of the models is calculated using Equation (11).

Table 4. Total power loss and efficiency (i/p voltage 20 V).

No. of Phases	Power Loss (W)	Input Power (kW)	Output Power (kW)	Efficiency (%)
2	3.2	51	47	94
3	3.5	50	48	94.5
4	3.6	50	48.5	94
6	3.1	50.3	47.5	94.5

and

Table 5. Total power loss and efficiency (i/p voltage 200 V).

No. of Phases	Power Loss (W)	Input Power (kW)	Output Power (kW)	Efficiency (%)
2	268	6.2	5.9	96
3	266	6.3	6	96.5
4	267	6.3	6.5	96.7
6	270	6.5	6.8	96.78

provide detailed insights into the total power loss and efficiency corresponding to input voltages of 20 V and 200 V, respectively. It is evident that when the phase increases, there is a decrease in power loss and an improvement in % efficiency. With less power loss than the four phases of the model, the six-phase model achieves efficiencies of 94% and 96%, respectively, whether the input voltage is 20 V or 200 V. Lower power losses accompany this compared to the four-phase model. Equation (11) reveals that increasing phases decreases capacitor and inductor values. Compared to the three-phase, four-phase, and two-phase Interleaved Boost Converters (IBCs), the six-phase IBC demonstrates significantly reduced inductor and capacitor values. Consequently, this produces a substantially smaller physical size for both the inductor and capacitor components.

$$P_{input}=P_{output}+P_{diode}+P_{capacitor}+P_{inductor}+P_{diode}$$

(10)

$$Efficiency \eta \left(\%\right)=\left({\displaystyle \frac{P_{output}}{P_{input}}}\right)\ast 100$$

(11)

A comparison with the current converters, like BC and IBC, is performed to verify the results produced by the model. Passive parts like the capacitor and inductor are made to work with all the converters used as references. The six-phase converter is compared with the standard BC and IBC parameters. Table 5 clearly shows how the model provides a higher output voltage and power. The output voltage and capacity are 500 V and 34.38 kW, respectively. A 26.92% reduction in inductor current occurs. Significantly lower values of capacitors and inductors are achieved using the proposed model. For the six-phase model, the importance of the inductor and capacitor are 168 F and 66.67 H, respectively. As a result, it is concluded that the six-phase model is better than the existing converter. The proposed converter is contrasted with the one currently being used to validate the six phases. The comparative result shows that the suggested converter’s efficiency is greater than the current converter’s at 98.6%.

Table 6. Comparison of existing approaches.

Parameters	Standard Converter Model [24]	Traditional Converter Model [25]	Proposed Model with Routing Protocol
Voltage (I/P) (V)	200	200	200
Voltage (O/P) (V)	300	300	600
Total phases	1	5	7
Power (O/P) (kW)	20	25	35
Frequency (kHz)	30	15	25
Current (A)	200	200	185
$\mathrm{Inductor} (\mathsf{\mu }\mathrm{H}$)	350	50	70
$\mathrm{Capacitor} (\mathsf{\mu }\mathrm{H}$)	750	190	170
Duty cycle	0.5	0.6	0.5
Current Ripple (A)	13	13	10

presents this comparison in full, along with additional factors.

4.2. Discussion

The proposed model’s efficacy was validated by creating and assembling a simulation prototype designed explicitly for two-phase operation with switching frequencies of 5 kHz and 10 kHz. This prototype operates within a supply voltage range of 0 to 5 volts, utilizing a nominal voltage of 5 V to drive the converter. The output voltages at 5 and 10 kHz switching frequencies across different duty ratios are considered. Notably, as the load voltage increases, the duty ratio also experiences a corresponding rise. The evaluation of output current, voltage, and power at these two switching frequencies reveals that the load voltage reaches a higher 13.2 V at a switching frequency of 5 kHz. The output voltage at 10 kHz is 0.94 times lower than this particular output value. It is observed that with an increase in the supply voltage, the output current also rises, resulting in a greater output power at the 5 kHz switching frequency compared to the 10 kHz switching frequency.

Table 7. Comparison of existing with proposed approaches.

Parameters	Boost Converter [31]	Buck-Boost Approach [32]	Interleaved Boost Converter [12]	Fly Back Approach [37]	Proposed Model
Input voltage (V)	25	25	25	25	25
No. of phases	1	6	2	6	6
Switching frequency (kHz)	25	25	25	25	25
Duty cycle	0.6	0.6	0.6	0.7	0.6
Output voltage (V)	200	208	205	213.5	500
Output power (W)	450	453	460	487.6	343
% Efficiency	95	96	96	97	99

shows comparison of the efficiency performance with other related metrics of those proposed with traditional buck-boost and fly back methods. Based on the comparison, it is shown that the efficiency of the proposed model is 99%, which is higher than the buck-boost with 96% and 97% for fly back approaches, respectively, where the input voltage is 25 V, No. of phases is 6, switching frequency is 25 kHz and duty cycle is 0.6. Figure 6 depicts the anticipated efficiency measure of the proposed EV model under three successive distance metrics: 25 Km, 40 Km and 60 Km. The efficiency is plotted versus speed. The figure shows that better efficiency is achieved when the distance and speed are higher. The electric vehicle runs in a more stable and optimized range at higher speeds and longer distances, which reduces energy losses and improves overall efficiency. Similarly, Figure 7 shows the efficiency measure due to power transfer. The transmitted power plus accepted power and efficiency are analyzed with respect to speed that ranges from 80 to 150 km/h. It can be observed that the power conversion efficiency decreases with speed.

Figure 8 shows the packet delivery ratio (PDR) of the system versus total messages. It can be seen that the PDR is better at a higher transmission rate. Figure 9 shows that the throughput increases with respect to the number of EVs. The proposed methodology is consistently superior to traditional and standard DVMRP methods in terms of PDR. This shows that the proposed method is more effective in providing packet delivery under different kinds of network challenges, resulting in improved communication reliability. In comparison to standard and conventional DVMRP protocols, the suggested model first shows a greater E2E latency when sending 500 packets. Due to the additional processing and control overhead of the improved features in the proposed model, there has been an increase in delay. However, the suggested model performs better in terms of delays as network conditions change. The proposed approach gradually beats conventional methods in other circumstances, proving its capacity for better routing and minimization of latency over time.

Figure 10 shows the end-to-end delay of the proposed and existing systems. It can be seen that although the delay increases with the total number of packets, the proposed system offers the lowest delay. Finally, Figure 11 compares the anticipated efficiencies of the buck-boost, fly back and proposed models where the proposed system promises the highest efficiency. Under certain situations, the proposed Enhanced DVMRP protocol might have higher delays initially; however, in in the long term, it offers more effective PDR and efficiency performance. As the system becomes stable, it also achieves reduced delays in different kinds of situations, showing its effectiveness in maximizing network efficiency and communication.

The proposed model not only enhances operating efficiency but also helps to reduce the total carbon footprint. This technique helps in reducing the amount of energy required for charging and operation by optimizing energy consumption and enhancing power transfer efficiency. This reduction in energy use directly results in a lower carbon footprint, especially when the electric power used for charging is obtained from renewable sources.

5. Conclusions

EVs require fewer natural resources to make, which opens the door for using renewable materials and cutting down on resource depletion. If EVs are charged with electricity from renewable sources, it can significantly reduce gas emissions and their impact on climate change. Engineers continuously leverage emerging technologies to enhance power systems, focusing on integrating the Internet of Things (IoT) into practical applications, especially within the energy industry. This research proposes a real-time solution that utilizes advanced vehicle-to-charging station (CS) communication through modern technologies to optimize electric vehicle (EV) charging schedules. The suggested approach introduces enhanced routing with DVMRP, offering advantages such as restricted aggregator access to end-user information, preventing unnecessary congestion at charging stations, user-friendly interfaces, access to comprehensive data about nearby CSs, and flexibility to select charging times or costs that align with their preferences. This robust tool not only aids CSs in anticipating future loads, preventing congestion, and optimizing energy usage, but also contributes to load balancing for EV customers. The proposed method has a remarkable efficiency of 99%, which is 2% higher than existing converter models, substantially reducing energy losses and making the proposed design highly effective for EV applications. This approach is superior to traditional converters such as boost converters, buck-boost converter, interleaved converter and fly back models, providing smooth speed control, which becomes essential for maintaining optimal driving conditions in electric vehicles. By increasing efficiency and minimizing energy losses, the proposed converter helps to extend battery life, making EVs more reliable and cost-effective over time. The high efficiency and reduced energy losses also contribute to a reduced ecological footprint, which is consistent with global sustainability goals. The ability to integrate the proposed converter with renewable energy sources expands its usefulness in green energy systems. This converter provides better performance, smooth speed control, and longer battery life.

In future developments, integrating vehicle-to-vehicle (V2V) and vehicle-to-grid (V2G) functionalities can further solidify grid reliability. The traditional notion that every device must communicate directly with the grid operator has posed challenges to deploying smarter grids. To overcome this, the research suggests a more efficient approach where devices connect to a dedicated web page, allowing configuration to respond optimally to the grid operator’s demands and load requirements. While this decentralized approach improves efficiency, it is acknowledged that this solution’s computational burden and complexities may grow significantly as the number of EVs increases. Additionally, the affordability of mobile networks may lead to overcrowding, potentially impacting communication quality. Addressing these challenges will be crucial for the successful implementation and scalability of the proposed solution.