One of the big challenges related with electric vehicle (EV) market penetration is the charging process, where the main problems are related to the lack of proper infrastructure in residential buildings (condominiums) since they are not prepared for this new reality. Condominiums have the problem of shared electricity, which does not meet the EV owner’s requirements. Based on new advances in the Internet of Things (IoT) [1
], and the associated sensing devices and communication platforms, blockchain and information systems have the potential to create new solutions for these problems. Another facet of this challenge is the problem associated with rental houses and the eventual need for supporting EV charging in these cases.
In condominiums, unfortunately, there is a general reluctance regarding the installation of EV charging stations that will only be used by a few homeowners [2
]. In addition, there is also an issue regarding the safety of the electrical installations, since they are not built proactively to support EV charging stations, and, adapting the condominium electrical infrastructure will require not only that a consensus between the majority of the owners is reached, which may be hard to achieve, but also authorizations issued by the government building safety entities.
Taking into consideration that most residential buildings have shared spaces with common electrical installations and are not prepared for the installation of new EV charging systems, this is a barrier to EV uptake [3
]. A study by Lopez-Behar et al. [4
] identified four main problem domains in the context of sharing EV charging solutions in buildings: unavailable charging infrastructure, building limitations, regulation issues and parking availability.
In this work, we propose a new IoT-based approach for handling the EV charging process, which can be used in the context of a shared energy infrastructure without requiring a supervision entity to control the process.
The proposed solution is supported by a decentralized blockchain approach, running on a mobile device app. Figure 1
shows an overview of a condominium with the proposed EV charging platform. This work allows the following features: (1) A pre-registration with a local EV charging provider is not required, avoiding the problem of different cards in different charging infrastructures (every charging infrastructure has its own cards, and this is a problem for EV owners because they need several charging cards when different providers are available); (2) it can work with digital currency using a peer-to-peer (P2P) framework on the same homogeneous blockchain infrastructure and technology; and (3) reduced cost (almost zero fees), because there is no requirement for a third party management entity, apart from the condominium, which would create additional costs.
As illustrated in Figure 1
, the major features of the proposed system are: (1) User authentication with a mobile device using Bluetooth Low Energy (BLE) communication and, based on this, release of energy for the EV charging process; and (2) energy consumption is monitored by Internet of Things (IoT) sensors and a microcontroller board transmits the data to a web server (Raspberry Pi with Raspbian operating system), which acts as the management unit, storing the data, handling the transactions in a blockchain implementation and managing the charging according to the power limitations.
Complementary to the setup presented in Figure 1
, which is suitable for deploying the solution at the local level, in the context of a single condominium, an equivalent model can be applied to scale the solution to a wider geographical area with an increased number of charging locations. In this sense, Figure 2
expands the proposed model to an IoT architecture that is suitable to explore cloud paradigms, such as Infrastructure as a Service (IaaS) or Software as a Service (SaaS), where the local management unit is replaced by a shared cloud computing platform. Without loss of generality and instantiating the model with existing platforms, the mobile app can be deployed on the Google Play store or Apple’s App Store, the Management Unit can be packaged in a Docker container [5
], and deployed on the AWS (Amazon Web Services) cloud computing platform, and the Ethereum open blockchain network can be used to support the financial transactions originated by the EV charging operation.
also enumerates the sequence steps to initiate a charging process: (1) Using the internet connection, the payment is sent from the mobile device to the open blockchain network (Ethereum); (2) the information related to the operation is exchanged between the mobile device and the Management Unit hosted on the AWS; (3) payment is received from the blockchain network, triggering the charging process on the Management Unit; and (4) the EV charging process is enabled on the IoT device (installed on the parking facilities), and the information related to the energy being delivered is sent back to the Management Unit on the AWS.
This paper is organized as follows. Section 2
presents the state of the art in related work. An overview of the proposed approach is presented in Section 3
, and Section 4
describes the system implementation. Section 5
presents a case study at a condominium, and Section 6
discusses future implications of the presented work. Finally, Section 7
presents the conclusions.
2. State of the Art
The proposed approach explores a set of works in several domain areas to create a new approach to handle the EV charging process in shared spaces, including the use of IoT sensing information to measure electricity taken on the EV charging process. Concerning driver profiles and EV charging with power limitations, several studies have been performed, and we apply an approach based on our previous work described in [6
]. In our implementation, it was also considered an implicit authentication mechanism [7
], applied on user’s mobile devices, which confirms the user authentication based on actions that he had performed on a daily basis. This implicit authentication mechanism can be used to prevent fraudulent credit transactions on a mobile device, verifying that the user is who he claims to be during the transaction. After researching systems that meet our criteria, we found some promising work [8
]. We apply a solution with user privacy (no identification is performed) in an approach based on the system proposed by Frank et al., called Touchalytics [13
]. We also apply the blockchain approach to handle distributed transactions without central supervision. The primary goal of the blockchain is to allow decentralized transactions with a digital currency, such as Bitcoin [14
] or Etherum [15
], without the need of a public authority to control the process. From the technical perspective, a blockchain is a sequence of blocks associated with transactional data using encryption based on a private and public key [16
]. User A
performs a transaction, and this process is associated with a block encrypted with his private key, in a hash process. User B
checks the transaction using the public key of user A
, allowing the following properties:
Decentralization, since we need confirmation from some party of each block transaction without central control;
Anonymity, since it allows for the authentication of transactions without giving up any personal information;
Auditability, which is performed based on the fact that each of the transactions is recorded and validated with a timestamp, where users can trace the previous transactions by accessing any node in the distributed network.
The application of blockchain in the domain of smart grids has great potential, providing a decentralized approach to implement management systems [17
] and handle power transactions. Due to the large space occupied by the meter sampling information on a blockchain block, [17
] presents a design to balance the amount of information kept onchain/offchain while keeping the properties of a block chain implementation. The authors of [18
] note the use of an open public cryptocurrency network, such as Bitcoin or Ethereum, can introduce a high transactional cost, due to the fees associated with cryptocurrency transaction processing (eventually similar to the cost of the energy supplied), and propose the development of a private Bitcoin-based blockchain network for EV charging purposes. Other relevant application cases include micro-generation [19
], as well as the contribution to handle the EV charging payment process without the use of propriety company payment systems.
The EV charging payment process is more frequent than fossil fuel refuelling and more complex due to the immaturity of the service. Specifically, the following issues are fairly common: (1) Transparency and clarity of rates and charges before they are incurred; (2) ability to pick-and-choose best rates and location of available charging points on the go; (3) ability to request priority charging and pay for it, when other EVs do not need priority; (4) ability to select a supplier or source of electricity, which would also enable greater competition and increase trust of customers; and (5) preferences for various types of payment, such as post-paid, pre-paid, or one-off payment.
We complement this work with our previous work on an EV charging system [21
] and IoT energy measurements using local sensors [23
], as well as new challenges of energy markets [19
]. Some issues identified are also addressed in [24
], which proposes a blockchain-based model with recourse to a bid to identify charging stations (and eventually schedule the charging), complementary to the approach suggested in [21
]. Another issue originated by the increase of the EV charging needs is the impact on the energy demands and the power limitation of the existing infrastructure [25
], which may not only increase the operational costs to fulfil the required demand, but also affects the voltage stability of the network. In [25
], the authors introduced the AdBEV, which is an algorithm to optimize the EV charging schedule, maximizing the voltage stability at the power grid side, and minimizing the charging costs. In [26
] the application of a blockchain-based process is suggested to support the EV charging queue management.
Together with mobile device authentication and a payment system, we developed a new approach to be used in shared EV charging spaces. Another interesting output is to use mobile devices to provide authentication and payment services in the context of the public EV charging systems, exploring recent advances in mobile device payment systems for public transportation [27
] and other application areas [28
]. As a new topic of research, new publications are appearing in the literature concerning the use of a blockchain approach to handling the EV charging process, such as: testing pilots to use digital currency for the EV charging process [29
]; proposal of a P2P energy transaction model to handle the EV vehicle-to-grid (V2G) operation in smart grids [31
]; handling the EV authentication issues based on a blockchain approach [32
]; proposal of a cross-domain authentication scheme with blockchain [33
]; and handling of security and privacy issues for energy transactions based on blockchain. Moreover, in this context, the EV is identified as part of the energy market [34
], and as a contribution to the contextualization of the local energy market [35
], where the blockchain plays an important role in the decentralization process, as well as for optimization purposes [36
5. Case Study at a Condominium
We applied the current approach to a shared place in a condominium, where three EV owners shared the condominium electric installation available at parking places for a period of 3.5 months. Each sensor was configured to generate one sample each minute, allowing further study of the current load patterns during a charging event. A set of three EVs (all Leaf vehicles with 24 kWh battery capacity) and three independent sensors (Sensor 0; Sensor 1; Sensor 2) were used; Figure 17
presents the diagram of the test environment for the case study. Due to physical constraints of the installation, the charging adapter connected to Sensor 0 was directly connected to the power grid, without one intermediate switch (“always on” on the scheme).
presents photos of one of the developed prototypes. Figure 18
a shows a photo of one of the IoT unit prototypes installed (Label (3) in Figure 18
a) of the test environment, measuring the current with the non-intrusive SCT-013-000-100A sensor (Label (1) in Figure 18
a). In this case, due to the weak Wi-Fi signal at the install location and the absence of other network infrastructure, the sensor unit was connected, using the RJ45 Ethernet interface, to a Wi-Fi Range Extender (Label (2) in Figure 18
a) to amplify the signal, allowing the IoT unit to reach the Management Unit accessible from the network where the Wi-Fi Range Extender was connected. Figure 18
b shows the contents of the IoT unit installed in Figure 18
a (Label (3)).
summarizes the data collected during the case study.
shows the charging time and the average charging power for each charging event (for events with > 3 h of charging duration), where it is possible to identify an average value of 2.3 kW, approximately (assuming an root mean square, RMS, voltage value of 230 V). The absence of a strong correlation between the charging time and the average charging power is also observable (the correlation coefficient between the charging duration and the charging power dataset is −0.30), which suggests that the average charged power by hour load is limited by the charging device and not directly dependent of the amount of energy required to charge the EV (e.g., a charging event of 6 h has a similar average charging power as a charging event with 3 h).
displays simultaneous charging events for the entire period analysed (330 charging events on 20 January and 12 May). Due to the power limitations, only two EVs are allowed to be charging at the same time, using full charging power, and the power is delivered on a first-come, first-served (FCFS) basis, where the platform controls the maximum number of stations that are allowed to charge the EVs simultaneously, queueing the remaining charging requests until a charging slot is available.
Since the charging platform measures the supplied power continuously, it detects when the EV is fully charged. At that time, it interrupts the EV charging process, transfers the data to the blockchain network (to account for the transaction performed), and starts supplying energy to the next EV queued. Supported by the drivers’ consumption profile and the statistical information about their behavior (taken from past stored data, average power required, the average number of hours before the vehicle is unplugged, etc.) a priority/utility-based resource scheduler can be applied to maximize the benefits/utility of the energy supplied.
shows the charging sessions of a Leaf with 24 kWh battery capacity, in a 3.5 month period, where it is possible to verify charging session periods ranging from 1 to 9 h (with an arithmetic average of 5.12 h and standard deviation of 2.03 h), and Figure 22
shows the charged energy, which varies between 2 kWh and 22 kWh (with an arithmetic average of 11.67 kWh and standard deviation of 4.58 kWh). It is possible to identify in this figure that, on average, this driver only charges 52% of the total charge and uses, on average, 5.5 h to charge the EV. From this approach, it is possible to identify driver profiles and use this for future charging processes accounting for the power limitation, as is shown in [19
shows the charging process with three EVs at the condominium, where it is possible to identify that, due to the power limitation, EV2 had to wait for an available charging window.
presents the distributions for the charging time (left) and for the charged energy (right) for each charging event. It can be observed that for 89% of the charging events ((117 + 82 + 69)/300), the EV will be charging for 6 h or less. A similar analysis can be made for the charging energy, where for 92% ((108 + 93 + 76)/300) of the charging events, the EV will charge 15 kWh, which represents roughly 62.5% of the total battery capacity.
Several usages pattern also were observed. Figure 25
displays the distribution of the amount of time between each EV charging event, which shows that for 64% of the times the driver charges the EV with less than 20 h between charging events, which may be correlated with the commute journey.
The work presented in this paper explores different approaches based on IoT, mobile devices and blockchain to create a novel solution for the EV charging process in shared spaces with authentication and security features, accounts and a transaction system. This approach can contribute to the proliferation of EVs, because one of their current barriers is the charging process at condominiums and rented houses. Moreover, from this solution, it is possible to identify EV charging profiles, create patterns to handle power limitations and share services without the need for new individual services. This approach can also be applied to handle energy transactions in other application scenarios, such as micro-generation without a central supervision control mechanism, although the use of open public cryptocurrency platforms like Bitcoin or Ethereum, due to high transaction costs, can create some barriers to the acceptance of the model.
The proposed solution demonstrated the robustness of the developed prototype for an EV charging process in shared spaces in the context of the presented case study at a condominium. During the 3.5 month of operation, there was only one failure of an IoT sensor unit due to a general power failure, and the problem was corrected by simply delaying the start of the charging process. Although no network-related limitations were identified while using traditional wired (Ethernet) and wireless (Wi-Fi) local area network (LAN) technologies to establish communication between the IoT devices and the Management Unit for the presented case study environment, the implementation of the system in wider geographical environments or other building topologies may require the use of wireless communication technologies more suitable for that context, for instance, low-power wide-area network (LPWAN) technologies such as LoRa, Sigfox, NB-IoT or LTE-M.