Received 28 January 2023, accepted 9 February 2023, date of publication 13 February 2023, date of current version 23 February 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3244562

A Techno-Social Approach to Unlocking Vehicle

to Everything (V2X) Integration:
A Real-World Demonstration
SHIVAM SAXENA 1,2 , (Student Member, IEEE), HANY E. Z. FARAG 1, (Senior Member, IEEE),
LEIGH ST. HILAIRE3 , (Member, IEEE), AND AIDAN BROOKSON3
1 Department of Electrical Engineering and Computer Science, York University, Toronto, ON M3J 1P3, Canada
2 Hero Energy and Engineering, Vaughan, ON L3T 0C5, Canada
3 Volta Research Inc., Toronto, ON M6P 0B3, Canada
Corresponding author: Shivam Saxena (saxenas@[Link])
This work was supported by the Independent Electricity System Operator (IESO) Grid Innovation Fund.

ABSTRACT The urgent transition towards electric vehicles (EVs) may potentially result in expensive
capacity upgrades for power system operators (PSOs), however, the recent uptake in bidirectional EV
charging, often referred to as ‘‘Vehicle to Everything’’ (V2X), may defer this need while simultaneously
generating revenue for participating EV owners. Yet, there is a lack of real-world investigation to elicit (a) the
willingness of EV owners to participate in V2X programs (b) the alignment of these programs within existing
electricity markets, and (c) how V2X may be combined with other distributed energy resources (DERs) to
realize maximum energy savings. As such, this paper proposes the design and implementation of a V2X
program that is informed by the socio-technical preferences of EV owners, including convenience, revenue,
and emissions savings. A survey is distributed to 124 EV owners to characterize their degree of willingness
to participate in the program, which leads to the design of several EV session types that enable EV owners
to arbitrage at their convenience, or participate in demand response (DR) events for higher revenue. Real-
world tests at residential and commercial buildings in Ontario successfully demonstrate the facilitation of
both arbitrage and DR, leading to increased revenue, decreased emissions, and the ability to participate in
DR with other building DERs to reduce the building load by 30 kW. The proposed work contributes a novel
attempt to demonstrate pathways to realize the potential offerings of V2X programs to benefit both EV
owners, building owners, and PSOs.

INDEX TERMS Bidirectional charging, demand response, distributed energy resources, electric vehicles,
grid services, microgrid, vehicle to everything, vehicle to grid, transactive energy.

I. INTRODUCTION systems, leading to load congestion, voltage and frequency

The urgency of transitioning towards electric vehicles (EV) deviation, as well as extended power outages [2]. Of partic-
has accelerated in recent times, with more than 20 countries ular importance is the scenario where large numbers of EVs
mandating the exclusive sale of EVs by 2030 [1]. While charge their batteries at the same time, which may coincide
this transition will significantly reduce tailpipe greenhouse with pre-existing peak times of electricity, and would require
gas emissions (GHGs), there is growing concern that the expensive and time-consuming grid capacity upgrades inclu-
resultant rise in electrical demand may overburden power sive of distribution transformers, feeders, and substations [3].
With the issue of rising peak load in mind, recent inno-
The associate editor coordinating the review of this manuscript and vations in EV technology have aimed to turn the EV into a
approving it for publication was Akshay Kumar Saha . bidirectional source of energy that uses its on-board battery

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see [Link]
VOLUME 11, 2023 17085
 S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

as an auxiliary power source [4]. The pervasive outlook for V2H-enabled scheduling algorithm to minimize the time a
this technology has led academia and industry to coin a new home spends without power.
name, ‘‘Vehicle to Everything-X’’ (V2X), where power from Building on the two previous criteria, the works in [17],
the EV can be used by electrical loads directly plugged into [18], [19], [20], [21], and [22] explore the aggregation of
the EV (V2L), by homes and buildings via a bidirectional V2G to provide grid services, including the modulation of
EV charging station (V2H/V2B), as well as the greater elec- reactive power for voltage regulation at the distribution level
trical grid (V2G) [5]. Apart from peak load reduction, V2X [18], [19], as well as active power modulation for frequency
can also provide other benefits at both the distribution and regulation at the transmission level. Similarly, the works in
transmission level, including voltage/frequency regulation [17] and [20] use rolling-prediction based frameworks to
and renewable energy firming [6]. It is worthwhile to note estimate the total capacity that could be offered into capacity
that the term V2X is also commonly applied within commu- markets, finding that residential V2G could avoid capacity
nication paradigms involving vehicle to vehicle (V2V) and upgrades needed for full electrification of light-duty EVs.
vehicle to infrastructure (V2I) [7], however, without loss of The authors in [21] utilize commercial EVs to provide load
generality, this manuscript will refer to V2X in its collective variance minimization and voltage regulation, which uses a
capacity to deliver energy within power grids, with specific fuzzy-decision-making predictive control strategy to target
mention of the destination of the energy flow described when load variances and discharge EVs at strategic times. Further-
appropriate, whether V2L, V2H, V2B, or V2G. more, the work in [22] proposes a multi-market optimization
The topic of V2X has gained a great amount of interest model that enables EVs to participate simultaneously in both
in recent times, where researchers have made strong contri- energy and ancillary markets, leading to increased revenue
butions. A tabular summary of recent research in this vein and decreased charging infrastructure costs.
can be found in Table 1, while a more detailed discussion is Although prior research has contributed tremendously to
presented hereunder. the state-of-the-art in V2X, several aspects are preventing its
The works in [6], [8], [9], [10], and [11] characterize EV improved uptake. While several works have considered EV
owner preferences, such as anticipated departure time and owner preferences in the design of their proposed algorithms,
discomfort caused by range anxiety, into parameters that these preferences have not been informed by real-world sur-
define the total usable capacity of the EV for discharging veys or data, nor is it clear to which degree the preferences
purposes. In [6], an optimal scheduling algorithm is proposed would affect the EV owner’s participation in V2X. For exam-
for EVs to provide voltage and frequency support to the ple, the considered preferences focus mainly on monetary
transmission system in consideration of battery degradation gain, and do not reflect environmental or societal benefits
tolerance and anticipated time of departure. Furthermore, that could play a major role in the decision to participate in
in [8], a deep reinforcement learning algorithm is proposed to V2X [23]. Moreover, adequate steps have not been taken to
consistently balance EV owner range anxiety and preferred reduce the complexity of V2X and improve its awareness
charging location with financial gain under highly dynamic for all parties, including power system operators (PSOs),
pricing. The work in [9] investigates the problem of V2G EV owners, and energy aggregators, making it difficult to
shifting peak load considering the randomness of EV owners design programs that could maximize benefits for all in an
arrival and departure times, and proposes a multi-objective intuitive manner [24]. In addition, V2X could play a signif-
satisfying platform that delivers appreciable load shifting and icant role by being aggregated with other DERs and partici-
daily EV operation savings of over 30%. The authors in [10] pating in grid services, however, its realistic alignment with
develop a distribution system plan to minimize grid capacity current electricity market rules, including the governance and
upgrades in view of locational preferences for EV public contracting between all parties, is left unclear [25]. This is
charging and discharging, while in [11], the authors propose an important oversight as PSOs begin to test the ability of
a distributed, multi-objective control solution to minimize mixed aggregations to perform in capacities as low as 100 kW
battery degradation and maximize daily EV owner revenue [26]. Lastly, there is a lack of real-world V2X deployment and
under real time pricing incentives. lessons learned to date, which is critical in establishing trust,
On the other hand, the research in [12], [13], [14], [15], market readiness and eventual uptake. It is worth noting that
and [16] focuses on integrating V2X with other DERs for the papers described in Table 1 do not use real-world data sets
optimizing power flow in behind the meter or off-grid appli- within their work, thus speaking to the inaccessibility of this
cations. In [12], [13], and [15], optimal dispatch algorithms data.
are proposed to aggregate wind, solar, and stationary energy As such, this paper proposes the design, development, and
storage systems with V2B to reduce the peak load of buildings deployment of a V2X program that is tested at real-world
and/or arbitrage according to electricity prices. The authors residential and commercial buildings in Markham, Ontario.
in [14] propose a cooperative power sharing strategy to shave To investigate often neglected social aspects of V2X, a survey
peak load between four interconnected microgrids in concert is distributed to 124 EV owners to characterize concerns,
with solar energy, leading to reductions of peak load between benefits, and criteria to V2X participation, thus establishing
67% and 90%. On the other hand, the work in [16] develops a baseline preferences to design the V2X program. Specifically,

17086 VOLUME 11, 2023

S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

TABLE 1. Comparison table of related and proposed work.

FIGURE 2. Summary of EV Owner survey.

These attributes are then parameterized into metrics that are

useful in observing trends that establish preference hetero-
geneity and influence the overall willingness to participate,
which informs the design of the proposed V2X program.
FIGURE 1. Contributions of proposed work.
As seen in Figure 2, the survey is divided into three sections
that ascend in order of comprehension complexity due to
the general unfamiliarity of EV owners with respect to V2X
the survey seeks to measure the financial, environmental, and [23], [24]. The first section focuses on asking general ques-
societal incentives required to secure EV owner participation, tions that establish behavioral patterns of the EV owner, such
combined with the willingness to leverage their EV for the as their preferred charging location, charging frequency per
provision of grid services such as demand response (DR). week, and estimated daily plug-in-time. The responses can
Based on the survey responses, a V2X program is developed be interpreted to ascertain the opportunity for EV owners to
to cater to the elicited preferences, which offer several EV participate in V2X, and predict the aggregated capacity (kW)
session types that enable an EV owner to select between and energy (kWh) available to participate in DR.
arbitrage or contracted grid support during DR events. The The second section gently introduces the concept of V2X
grid support mode is then aligned with the current DR market and describes the opportunity to earn incentives for participat-
in Ontario, where contracting and governance between EV ing in exchange for energy discharged and a commitment to
owners, building owners, and PSOs conforms with market plug-in to a charging station for a minimum duration. Partic-
rules for compliance. Both EV session types are tested with a ipants are also educated about the environmental opportunity
10 kW bidirectional EV charger at a commercial building, and for V2X in offsetting energy generation from fossil fuel based
paired with a 30 kW solar and storage system to demonstrate power plants. Two important parameters are introduced to the
the mixed aggregation’s ability to provide DR. Thus, as sum- EV owner at this stage, which are the minimum driving range
marized by Figure 1, the main contributions of this paper (MDR), and minimum-plug-in-time (MPT) [27]. The MDR is
are (a) the usage of survey data to define new parameters to the minimum level of range that is guaranteed to be left on the
characterize V2X preferences; (b) the design of a new V2X EV at all times to allow it to get to its next destination, while
program with configurable EV session types that prioritizes MPT is the minimum number of hours an EV owner would
the needs of both EV owners, building owners and PSOs; and be required to plug in to a bidirectional charging station per
(c) demonstration from real-world deployment and alignment month.
with electricity markets to help accelerate the uptake of V2X. The third section focuses on using the MDR and MPT
The organization for the remainder of the paper is as fol- parameters to introduce two different V2X program types
lows. Section II provides an overview of the administered and assess their interest in each. The first program broadly
EV owner survey, including key takeaways that govern the resembles a ‘‘pay as you go’’ program based on energy arbi-
design of the V2X program. Section III introduces the design trage, where EV owners would charge their EV during times
of the proposed V2X program itself, while Section IV focuses when electricity prices are relatively inexpensive (off-peak
on the software implementation of the program. Finally, hours), and discharge their EV during times when electricity
section V presents real-world results, while Section VI is prices are comparatively expensive (peak hours). This form
reserved for conclusions and future work. of arbitrage may also result in a lower carbon footprint for
the EV owner, since generally, the fuel mix of electricity
II. OVERVIEW OF EV OWNER SURVEY generated at off-peak hours is cleaner as compared to peak
A. SURVEY DESIGN hours, where fossil-fuel based peaking power plants may
The main objective of the EV owner survey is to cap- be procured to meet additional demand [28]. This program
ture their willingness to participate in V2X, comprising of type prioritizes the convenience of the EV owner over their
environmental, societal, economic, and technical attributes. potential to earn incentives, since the EV owner may choose

VOLUME 11, 2023 17087

 S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

FIGURE 3. EV Owner survey results showing (a) participation due to financial incentive, (b) participation due to environmental benefit,
(c) minimum incentive to participate (d) preferred V2X program, (e) profiling based on MDR and (f) profiling based on enhanced MDR.

to initiate and abort an EV session without any penalty, and worthwhile to note that individual consent to use the sur-
does not need to commit to any sort of contract with an energy vey responses for the purposes of knowledge dissemination
aggregator or PSO. The second program is based on a fixed within this paper was obtained during the survey process
contract, where the EV owner would commit to a MDR and itself, where participants indicated their consent by having
MPT, and receive payments for energy discharged ($/kWh), an option to opt-out of the dissemination activities.A total of
as well as capacity payments for the time spent plugged in 68% of EVs were from Tesla, Chevrolet, and Nissan, with
($/kW). If a DR event is called, the EV owner is entitled Tesla accounting for 38% (models S, X, Y, 3), the Chevrolet
to enhanced payment for energy discharged during the DR Bolt at 19%, and the Nissan Leaf at 11%. Further observations
event. However, the EV owner forfeits all monthly incentives from the survey results can be seen in Figure 3, and are
if they do not fulfill their level of MPT, or disconnects from discussed below.
the charging station for a certain period of time during a DR
event. 1) GENERAL WILLINGNESS TO TRIAL V2X PROGRAMS
To make the programs more intuitive to understand, three
As seen in Figure 3(a), there appears to be a general willing-
separate profiles are created to characterize ranges of prefer-
ness to engage in a V2X program across respondents, with
ence for MDR and MPT, entitled conservative, moderate, and
only 4% rejecting participation outright, 49% participating
aggressive. Without loss of generality, it is assumed that a
for a minimum incentive, and 47% participating regardless
conservative EV owner sets the MDR as two times their daily
of incentives. When told that V2X had the potential to offset
commute and the MPT at 4 hours, a moderate EV owner sets
the use of fossil-fuel based generators, 80% of the respon-
the MDR as 1.5 times their daily commute and the MPT at
dents cited enhanced motivation to participate, as seen in
8 hours, while an aggressive EV owner sets the MDR at their
Figure 3(b). Interestingly, when queried with respect to the
daily commute and the MPT at 12 hours. Monetary amounts
minimum incentive required to participate in a V2X pro-
are attached to each of the profiles and varied to observe the
gram, as seen in Figure 3(c), dominant responses included
trade-off between convenience and desire for incentives by
year-round free charging for their EV (42%), while 35%
determining if any of the EV owners changed their desired
indicated that they would participate as long as they did not
profile type.
lose any money. Put together, the survey results indicate that
the survey respondents possess a general willingness to trial
B. DISCUSSION OF SURVEY RESULTS V2X programs, strongly motivated by environmental benefits
The selection criteria to participate in the survey was for the and financial incentives.
participants to reside in Ontario, Canada, and to have driven Qualitatively, participants were also asked the reasons they
an EV before conducting the survey. The survey was subse- would not participate in a V2X program, where the top
quently disseminated online to EV owners in the Canadian reason cited by 33% of participants was the fear of acceler-
province of Ontario for a period of 9 months between March ated battery degradation due to frequent charging/discharging
2021 and December 2021, and garnered 124 responses. It is cycles, while 28% cited trust issues with the technology itself

17088 VOLUME 11, 2023

S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

First, the roles of participants in an EV session transaction are

defined, including the EV owner, energy aggregator/building
owner, and/or the PSO, followed by a detailed description of
the logistics of the EV sessions themselves. Lastly, a sub-
section is dedicated to introducing how the proposed session
types would integrate within the existing electricity market in
Ontario, along with the governance and contracting required
between the transaction participants.

FIGURE 4. Summary of EV Session types. A. EV TRANSACTION PARTICIPANTS

The intent of the proposed EV program is to provide the
EV owner with flexibility to charge or discharge their EV
and allowing PSOs to remotely control their EV. These con- at different locations, whether at home, work, or at a public
cerns underscore the importance of designing and deploying charger. Depending on where the EV owner plugs in, the
real-world V2X programs to mitigate these concerns. transaction participants will differ. If the EV owner utilizes
a charger at home, then the transaction only involves the EV
2) PREFERENCE FOR V2X PROGRAM TYPES owner and the PSO, where it is assumed that the transaction
EV owners indicated a strong preference for a no contract price would equal the regular rate of electricity at that point in
type program as seen in Figure 3(d), with 78% voting for time, which EV owners would earn if discharging, and pay if
no contract, versus 22% voting for fixed contract. To fur- charging. This transaction is most conducive to jurisdictions
ther explore the motivation to participate as a function of where electricity prices vary as a function of time, such as
MDR, EV owners were asked to choose between profile time-of-use. On the other hand, if the EV owner utilizes a
types of conservative, moderate, and aggressive, as discussed charger at work or in public, the transaction participants are
previously. As seen in Figure 3(e), 55% chose the moderate the EV owner and the building owner, with the transaction
profile, while 37% chose the conservative profile, with only price being set by the building owner.
8% choosing an aggressive profile. Next, financial incentives However, if a DR event is called by the PSO, they are
were attached to the profiles in two stages, where in the included in the transaction as they are distributing incentives
first stage, the annual payout for conservatives, moderates, to the EV owners and building owners for participating in the
and aggressives was $500, $750, and $1000 respectively, event. As such, it is assumed that the building owner would
while in the second stage, the annual payouts were enhanced act as an aggregator of EVs and other on-site DERs, and
to $750, $1000, and $1500, respectively. At the first stage, operate as the point of contact between EV owners and PSOs.
the aggressive profile reported an increase of 4 EV owners, This practice aligns with the direction of current and future
or 3.3% of the total respondents. In the second stage, shown in electricity markets, where aggregators procure DERs that are
Figure 3(f), the trend was observed again, where the number behind the meter and sell capacity and energy to upstream
of EV owners switching to the aggressive profile increased markets [29]. Thus, incentives would flow from the PSO
by 7 (or 6% of the total respondents). By increasing the annual to the building owner (aggregator), and from the building
payout for the aggressive profile by a factor of 1.5, the number owner to the EV owners. The contracting and governance of
of EV owners willing to lower their MDR and offer more these transactions will be discussed later in this section, when
capacity for grid services more than doubled, from 10 to 21. alignment with electricity markets is discussed.
In totality, it can be established that the most influential
characteristic that drives an EV owner’s willingness to partic- B. EV SESSION TYPES
ipate in V2X is their convenience, as evidenced by the strong Based on the preferences discovered from the survey that
preference for the no contract option. However, the survey prioritize both EV owner convenience as well as opportu-
evidence also suggests that both environmental benefits and nities to earn revenue via V2X, three types of EV sessions
financial incentives do play a secondary role in establishing are proposed, which are Charge My EV, Quick Support and
participation. Therefore, it follows that a V2X program must Extended Support (Figure 4). Charge My EV is provided for
provide options that cater to these preferences in an intuitive an EV owner that wishes to charge their EV as quickly as
manner, which includes facilitating both fixed contract and possible, and is designed to cater to those use cases where the
no contract options. EV owner is short of time, and may not be at their preferred
charging location. As such, Charge My EV accepts a target
III. DESIGN OF PROPOSED V2X PROGRAM battery level as a parameter and immediately dispatches the
This section introduces the design of EV session types within charger at the maximum acceptable power rate. Optionally,
the proposed program that cater to the preferences garnered a departure time can be added as a parameter for those EV
from the survey results discussed in the previous section, owners wishing to average their power consumption over a
which are summarized in Figure 4 as a quick reference. fixed duration of time. For discharging the EV, EV owners

VOLUME 11, 2023 17089

 S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

leverage the Quick Support or Extended Support session type.

In Quick Support, the EV owner provides their MDR, and
the charger is either dispatched at the maximum acceptable
discharging rate, or at an average discharge rate if a departure
time is provided. If a DR event is called during a Quick
Support session, the EV owner may ‘‘opt-in’’ to the DR event,
FIGURE 5. Participant contracting for demand response.
and they will receive an additional DR payment (in $/kWh)
from the PSO on top of the regular electricity rate. This use
case caters to the dominant socio-economic preferences of
on-site DERs. As an example for the proposed integration, the
EV owners found in the survey by (a) enabling revenue oppor-
market rules from Ontario’s DR market are used hereunder.
tunities by discharging at peak electricity rates; (b) lowering
Ontario’s transmission system operator, the Independent
carbon footprint by discharging to offset the use of peaking
Electricity System Operator (IESO), administers a DR mar-
power plants; and (c) maximizing EV owner convenience by
ket that contracts market participants, including aggregators,
not enforcing a penalty for aborting their EV session. In the
to reduce energy consumption during periods of high demand
context of charging location and V2X scenarios, it is worth
in exchange for economic incentives [30]. Market partici-
noting that the three session types encompass all forms of
pants bid into an annual capacity auction, committing to a
V2X within this paper, whether V2L, V2B, V2H, or V2G.
specific reduction of energy consumption, in kW, for each
Conversely, Extended Support is contract oriented, where
applicable hour. With commitments in hand, the IESO issues
the EV owner would establish their constraints as per MDR
a ‘‘stand by’’ notice to market participants by 7:00 on the
and MPT, which can be updated monthly. For each session,
morning of the day to begin preparing for a DR event, and
two additional inputs are taken into consideration, which are a
further issues an activation notice at least 2 hours in advance
maximum battery level, and an expected departure time. The
of the event itself. In context of this paper, the formulation of
expected departure time indicates the availability of the EV
the DR capacity bid relies heavily on the contracted capacity
to discharge should a DR event occur. In a typical Extended
between the building owners and EV owners, along with
Support session, the EV would be charged to the prescribed
the useable capacities of its on-site DERs. The average con-
maximum battery level and sit idle, waiting for a DR event
tracted capacity from EV owners opted-in to the Extended
to occur, where each hour plugged in to the station counts
Support session type is analogous to the average useable
towards the monthly MPT. If a DR event occurs, the EV
capacity of the EV, in kW, as a function of MDR and MPT,
discharges until its MDR is reached. In this case, the EV
and can be formulated as follows:
owner receives capacity payments for the entire time plugged
in ($-kW), on top of enhanced DR payments for the energy (STREV EV EV
j − MDRj ) ∗ CAPj
discharged during the session ($/kWh). However, if the EV UCAPEV
j = (1)
MPTjEV /NWD,j
owner aborts the session and remains disconnected for certain
number of minutes during an DR event, all payments for the where, j is an iterator for the number of EVs assigned to
month are forfeited. This particular session type caters to the building and operating in Extended Support session type,
those EV owners in the survey that aggressively seek revenue UCAPEV j is the average useable capacity of an EV in kW,
opportunities and do not mind fixing their schedule to realize EV
STRj is the average starting range of the EV in % (typically
them. at the beginning of a workday), MDR is the minimum driving
range in %, CAPEV j is the capacity of the EV battery in kWh,
MPT is the monthly minimum plug in time in hours, and
C. PROPOSED INTEGRATION WITH DEMAND RESPONSE NWD,j is the average number of workdays the EV is parked
MARKET IN ONTARIO at the building per month. From (1), the building owner can
As discussed in the previous subsection, aggregators act as the obtain the DR bid as follows:
point of contact between providers of DR and the PSO, and so
within context of this paper, we assume the aggregator to be NEV NX
DER
a building owner with a parking lot of bidirectional chargers,
X
DRBID
i = UCAPEV
i,j + UCAPDER
i,k (2)
as well as on-site DERs. Thus, it follows that the building j=1 k=1
owner will bundle capacity from parked EVs in Extended
Support mode alongside capacity from its on-site DERs to where, DRBID
i is the cumulative DR bid, in kW, entered by the
provide DR services within a PSO administered market. building owner into the DR market for market interval i, NEV
As seen in Figure 5, this would require contracting between is the number of EVs, k is an iterator for the number of on-
EV owners and the building owner to establish the available site DERs, NDER is the total number of DERs, and UCAPDER i,k
capacity from the EV owners as a function of individual is the useable capacity of the aforementioned DERs. It is
MDRs and MPTs, and a subsequent contract between the worthwhile to note that the useable capacity of each DER is
building owner and the PSO that culminates in a DR capac- a function of each unique type of DER and its constraints,
ity bid that bundles the contracted EV capacity with other whether energy storage, intermittent renewable, or distributed

17090 VOLUME 11, 2023

S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

generator. More information in this regard can be found

in [29].
Building owners are subject to non-performance charges if
the DR capacity is not delivered during DR events, where two
important charges are the dispatch charge, and the capacity
charge [30]. The dispatch charge states that the resources
providing DR capacity must stay within a percentage toler-
ance of their set-point for each five-minute interval during
a DR event, which can be formulated in our scenario as the
following constraint:
(1 − α)DRBID
i ≤ PAGG
i ≤ (1 + α)DRBID
i (3)
where PAGG,i is the average aggregate power generated by
the building owner, in kW, during interval i, and α is a
FIGURE 6. Block diagram of implemented V2X program.
percentage tolerance that is between 0 and 1. On the other
hand, the capacity charge ensures that the committed DR
capacity actually reduces the load of the building, and thus is
a function of the hourly baseline load of the building [30]. The
violates this constraint, the building owner may levy a penalty
constraint for the capacity charge can be realized as follows:
on the EV owner by forfeiting all expected revenue during the
BLi − CONi ≤ ϵ ∗ DRBID,i − SCHi (4) EV session (capacity payment and energy payment).
where BLi is the average baseline load of the building during
interval i in kWh, CON is the averaged measured consump- IV. IMPLEMENTATION OF PROPOSED V2X PROGRAM
tion of the building during interval i in kWh, ϵ is the percent- This section discusses the implementation of the proposed
age tolerance for the for the capacity charge between 0 and 1, V2X program in context of the EV session types and integra-
and SCH is a scheduled real-time constraint signal issued by tion with DR event handling discussed in the previous section,
the IESO, in kW. supported by a block diagram that shows the participants and
For the building owner to avoid capacity and dispatch dataflow in Figure 6. Central to the V2X program is a V2X
charges, the building owner would need to dispatch the EVs application, referred to as the V2X App, which acts as the
and on-site DERs according to control rules that satisfy the main interface between an EV owner and all other transaction
aforementioned constraints for every market interval. This participants. EV owners utilize this app to configure prefer-
begins with the calculation of the target aggregate dispatch ences, including session type and session configuration, and
setpoint (PTAR
i ): view measurements of their session in real-time. When at
home, the V2X app interfaces directly with the bidirectional
PTAR
i = DRBID
i − (BLi − CONi − SCHi ) (5)
charger and dispatches the EV with control setpoints accord-
The building owner may then attempt to meet the target set- ing to the EV session configuration, while receiving measure-
point by dispatching all EVs operating in Extended Support ments from the charger and transmitting them back to the
mode as follows: EV owner. The measurements provide details on the current
(CREV EV EV battery level of the EV, setpoint (charging or discharging),
i,j − MDRi,j ) ∗ CAPi,j
SPEV
i,j = (6) electricity price, and approximate time left in the session.
LDR
A PSO is able to signal a DR event by sending a notification
where SPEVi,j is the setpoint for every EV during each interval to the V2X app, which, in turn, requests the EV owner to
i, CR is the current range, in %, and LDR is the length of remain plugged into the charger for the duration of the event.
the DR event remaining in hours. If the sum of the available Depending on the response and the session configuration, the
EV capacity (SPEV i,j ) is less than the target aggregate dispatch V2X app will adjust the control setpoints to begin discharging
setpoint, other DERs will have to be dispatched to make up the EV appropriately.
the difference. On the other hand, for building owners hosting
After the DR event is complete, the building owner may bidirectional chargers and on-site DERs, a DER management
validate the performance of every connected EV operating in system (DERMS) is used for all DER control and coordina-
Extended Support mode using similar logic to the dispatch tion. An EV owner still utilizes the V2X app to configure
charge that the PSO uses to evaluate the building owner, preferences and session settings, however, these settings are
which can be formulated as follows: transmitted to the DERMS, which in turn, carries out the
(1 − α)SPEV local control appropriately. During a DR event, the PSO can
i,j ≤ Pi,j ≤ (1 + α)SPi,j
EV EV
(7)
send the DR signal to the building DERMS, which will then
where PEV
i,j is the average measured power from each EV at dispatch all EVs in Extended Support mode and on-site DERs
each interval, in kW. If the measured power from each EV as per the methodology discussed in the previous section.

VOLUME 11, 2023 17091

 S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

FIGURE 7. EV session visualization for grid support.

FIGURE 9. Single line diagram of building under test.

FIGURE 10. Snapshot of Hero DERMS dispatching DERs during DR event.

FIGURE 8. Photo of commercial building with BESS (left), solar (top), and
Nissan Leaf with bi-directional charger (right).
the standard AC Level 2 power transfer of the Nissan Leaf at
6.6 kW [31]. For the second case study, the experiments are
The V2X app is implemented as both a remote mobile/web conducted at the same commercial building, which addition-
application, as well as a standalone application on a tablet ally has a 6 kW solar array and 50 kWh BESS manufactured
that is meant to be hosted locally, near the charging station. by TROES Corp, as seen in Figure 8.
An example screenshot from the V2X app is seen in Figure 7, A single line diagram of the commercial building is
where the EV is currently operating in Extended Support shown in Figure 9, where in addition to the aforementioned
mode, with a configured MDR of 20% and departure time of DERs, two smart energy meters are deployed. The energy
15:30. Real-time values from the bi-directional charger can meters are manufactured by Acuvim, with the model num-
also be seen on the screen, with the EV discharging at nearly ber being Acuvim-II, which is revenue-grade and samples
10 kW, earning revenue at $0.20/kWh, currently at a battery measurements less than every 20 milliseconds. One meter
level of 86%, and with 2 hours and 34 minutes left until the is deployed at the building’s point of interconnection to the
session is complete. PSO-administered grid to measure the net power flow, while
one meter is deployed in front of all building DERs to mea-
V. RESULTS sure the DER net power flow. This way, the true building
This section discusses the results of two case studies that load can be calculated by subtracting these two measurements
demonstrate the proposed V2X program operating across from each other.
all three session types. The first case study demonstrates The DERMS platform used to control the building’s DERs
arbitrage in using a combination of Charge My EV and is entitled Harmonize, developed by Hero Energy and Engi-
Quick Support, while the second case study demonstrates the neering, which interacts with the DERs and energy meters
provision of DR using the EV in Extended Support mode using the Modbus protocol, as seen in Figure 10. The
in aggregation with solar and battery energy storage system DERMS also listens for simulated DR Event signals from the
(BESS). In the first case study, the EV is charged at a residen- PSO, which defines the event start time, end time, and values
tial building using standard Level 2 AC charging (maximum for the load that the building is expected to reduce based on
of 6.6 kW), and is discharged at a commercial building using its DR bid.
a Level 3 DC bidirectional EV charger (maximum of 30 kW,
but derated to 10 kW due to building constraints). The EV A. CASE STUDY 1: ARBITRAGE USING HOME
used is a Nissan Leaf 2019 SV with a battery capacity of CHARGING AND V2B
40 kWh, while the bidirectional EV charger is manufactured In Ontario, residential electricity prices follow a time-of-use
by Coritech. The impact of the DC Level 3 charger enables based schedule, with the summer off-peak periods being from
far more power to be discharged from the EV, as compared to 19:00 to 7:00 at 8.2c/kWh, the mid-peak periods being from

17092 VOLUME 11, 2023

S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

TABLE 2. Energy, emissions and cost metrics for arbitrage test. The daily metrics are summarized in Table 2, where, nega-
tive entries in the table indicate energy discharged, emissions
avoided, and revenue accrued. As seen in the table, over
the course of the five days, the EV owner had a net energy
balance of 7.7 kWh and contributed a net emissions savings
of 3.6 kgCO2eq, while earning a total of $12.63 and shifting
150.7 kWh of electricity consumption from the grid at on-
peak hours. While the earnings may seem relatively insignif-
icant, recall that 47% of EV owners from the survey data
were willing to engage in V2X without financial incentives
attached, and a further 49% of EV owners’ participation
depended on a minimum incentive, of which the most popular
answer was free EV charging, which was achieved during
the case study. Furthermore, the environmental advantages,
especially at scale, are promising due to the ability of V2X
to discharge at on-peak periods, which again aligns with the
survey data as 80% of EV owners cited enhanced motivation
to participate due to this very fact.

B. CASE STUDY 2: V2B AGGREGATION WITH

BUILDING DERs
This case study involves validating the ability of the building
FIGURE 11. Power vs price in V2B arbitrage test. to deliver contracted DR capacity of 30 kW over a 2 hour
duration, with percentage tolerances for dispatch and capacity
compliance set to 15% and 90%, respectively, as per [30].
The MDR for the EV is set to 20% of the battery state of
charge (SOC), while the minimum and maximum SOC levels
for the BESS are 5% and 95%, respectively. The DERMS
follows the priority dispatch discussed in Section III-C, where
first, EVs operating in Extended Mode (and engaging in V2B)
are dispatched, followed by other DERs. In this scenario, the
solar array is not dispatchable, so the BESS is set to dispatch
the remainder of the DER target after accounting for the
production of the EVs and solar array. In terms of revenue,
in addition to the electricity rate during the test $0.17/kWh,
the EV owner would receive an enhanced rate of $2/kWh as
FIGURE 12. Power vs emissions in V2B arbitrage test.
per [33].
The results of the test can be seen in Figures 13, 14, and 15.
7:00-11:00 and 17:00-19:00 at 11.3c/kWh, and the on-peak As shown in Figure 13, the building load stays fairly consis-
periods being from 11:00-17:00 at 7.0c/kWh. Correspond- tent throughout the test period, oscillating between 125 kW
ingly, the emission factor of the Ontario grid is significantly and 136 kW. The building load is also consistent with
higher during the on-peak periods, as the Ontario IESO uti- the building’s baseline consumption as seen in Figure 13,
lizes gas-fired generators to meet the increased demand [32]. which is approximately 131 kW. The DER contribution reacts
As such, this five day case study uses the Charge My EV accordingly, averaging 30.4 kW of load reduction during the
session to charge between the off-peak hours of 00:00-5:00, 2 hour period, and reducing the actual consumption of the
and uses the Quick Support session to engage in V2B by building by 60.7 kWh. The contribution of each DER can be
discharging during the on-peak hours of 11:00-17:00. The seen in Figure 14, with solar providing between 3 and 5 kW of
results can be seen in Figures 11 and 12, which plots the production, the EV providing approximately 10 kW, and the
EV power transfer against the electricity price ($/kWh) and BESS making up the rest to satisfy the DR capacity. In terms
hourly annual emission factor (gCO2eq/kWh) of the Ontario of meeting dispatch compliance, the dispatch tolerance of
electricity grid [32], respectively. In both figures, it can be 15% necessitates that the DER contribution during each
seen that the EV charges at approximately 6 kW when both 5 minute interval of the test must be greater than 25.5 kW,
the electricity price and emission factor are at their lowest, which was successfully achieved as seen in Figure 14. Mean-
while on the other hand, the EV discharges at approximately while, the capacity tolerance of 90% necessitated that the
5 kW when both the electricity price and emission factor are average of the difference between the baseline and actual
at their highest. consumption throughout the event be more than 27 kW, which

VOLUME 11, 2023 17093

 S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

with their PSO and building owner to provide DR with a

MPT and MDR. The Extended Support EV session type was
further aligned with the market rules for Ontario’s existing
DR program, with added consideration for aggregation with
other DERs. All three EV session types were validated at
real-world residential and commercial buildings, demonstrat-
ing the ability of the V2X program to deliver a 3.6 kg reduc-
tion of GHG emissions, EV owner revenue while arbitraging,
and participating in a DR event in concert with other DERs.
FIGURE 13. Metered values in 30kW/2HRS DR test. While the proposed V2X program can be generalized to
other regions and countries due to its flexibility in offer-
ing multiple EV session types that facilitate vehicle-grid
integration from home or public charging infrastructure, rec-
ommendations for future work revolve around the financial
feasibility of the program as it relates to electricity pric-
ing. Different regions of the world have different pricing
types, whether fixed, location-dependent, time-dependent,
or both [34]. With evidence from experimental results in this
paper suggesting that EV power generation can be reliable
and constant, the revenue received as a function of electricity
FIGURE 14. DER contribution in 30kW/2HRS DR test. pricing will be variable depending on the region. As such,
it is recommended that future work considers a feasibility
assessment for V2X within different worldwide regions to
determine the revenue pool available for EV owners to access,
which will eventually inform new pricing models for the
uptake s of V2X.

REFERENCES
[1] International Energy Agency. Global EV Outlook 2021. Accessed:
Jun. 2022. [Online]. Available: [Link]
outlook-2021
[2] M. S. Shamami, M. S. Alam, F. Ahmad, S. M. Shariff, I. AlSaidan, Y. Rafat,
FIGURE 15. State of Charge of EV and BESS during 30 kW/2HRS DR test. and M. S. J. Asghar, ‘‘Artificial intelligence-based performance optimiza-
tion of electric vehicle-to-home (V2H) energy management system,’’ SAE
was also successfully achieved as seen in Figure 13, for an Int. J. Sustain. Transp., Energy, Environ., Policy, vol. 1, no. 2, pp. 115–125,
Sep. 2020, doi: 10.4271/13-01-02-0007.
average reduction of 29.04 kW. It is worthwhile to note that [3] M. Aryanezhad, E. Ostadaghaee, and M. Joorabian, ‘‘Management and
the EV generated power at its maximum capacity of 10 kW coordination charging of smart park and V2G strategy based on Monte
throughout the test, and ended with a final SOC of 29%, Carlo algorithm,’’ in Proc. Smart Grid Conf. (SGC), Dec. 2014, pp. 1–8,
doi: 10.1109/SGC.2014.7090887.
which is well clear of the 20% MDR preference, as seen in
[4] M. Restrepo, J. Morris, M. Kazerani, and C. A. Cañizares, ‘‘Modeling
Figure 15. With the EV having discharged 19.71 kWh during and testing of a bidirectional smart charger for distribution system EV
the event, the EV owner earns $2.17/kWh as a combination integration,’’ IEEE Trans. Smart Grid, vol. 9, no. 1, pp. 152–162, Jan. 2018,
of the regular and enhanced rate, for a total of $42.8. doi: 10.1109/TSG.2016.2547178.
[5] V. T. Tran, M. R. Islam, K. M. Muttaqi, O. Farrok, M. R. Kiran,
and D. Sutanto, ‘‘A novel universal magnetic power plug to facili-
VI. CONCLUSION AND FUTURE WORK tate V2V/V2G/G2V/V2H connectivity for future grid infrastructure,’’
This paper proposed a V2X program designed to cater to the IEEE Trans. Ind. Appl., vol. 58, no. 1, pp. 951–961, Jan. 2022, doi:
10.1109/TIA.2021.3130106.
socio-technical preferences of EV owners, while also being [6] S.-A. Amamra and J. Marco, ‘‘Vehicle-to-grid aggregator to support power
able to be aggregated with other DERs and participate in grid and reduce electric vehicle charging cost,’’ IEEE Access, vol. 7,
DR. To characterize the willingness of EV owners to partic- pp. 178528–178538, 2019, doi: 10.1109/ACCESS.2019.2958664.
[7] X. Wen, J. Chen, Z. Hu, and Z. Lu, ‘‘A p-Opportunistic channel access
ipate in V2X, a survey was designed and delivered to elicit scheme for interference mitigation between V2V and V2I communica-
preferences with respect to MDR, environmental benefits, tions,’’ IEEE Internet Things J., vol. 7, no. 5, pp. 3706–3718, May 2020,
and economic reward. Two different types of programs were doi: 10.1109/JIOT.2020.2967647.
proposed in the survey, with a no contract option and a fixed [8] L. Yan, X. Chen, J. Zhou, Y. Chen, and J. Wen, ‘‘Deep reinforcement
learning for continuous electric vehicles charging control with dynamic
contract option, and from the results of 124 respondents, user behaviors,’’ IEEE Trans. Smart Grid, vol. 12, no. 6, pp. 5124–5134,
both programs received a significant amount of interest for Nov. 2021, doi: 10.1109/TSG.2021.3098298.
participation. Using the survey results, the V2X program was [9] W. Li, Z. Lin, H. Zhou, and G. Yan, ‘‘Multi-objective optimization for
cyber-physical-social systems: A case study of electric vehicles charg-
outfitted with three different EV Session types, enabling an ing and discharging,’’ IEEE Access, vol. 7, pp. 76754–76767, 2019, doi:
EV owner to arbitrage at their convenience, or fix a contract 10.1109/ACCESS.2019.2921716.

17094 VOLUME 11, 2023

S. Saxena et al.: Techno-Social Approach to Unlocking V2X Integration: A Real-World Demonstration

[10] X. Wang, Y. Nie, and K.-W.-E. Cheng, ‘‘Distribution system plan- [29] H. J. Kim, H. J. Kang, and M. K. Kim, ‘‘Data-driven bidding strategy for
ning considering stochastic EV penetration and V2G behavior,’’ IEEE DER aggregator based on gated recurrent unit–enhanced learning particle
Trans. Intell. Transp. Syst., vol. 21, no. 1, pp. 149–158, Jan. 2020, doi: swarm optimization,’’ IEEE Access, vol. 9, pp. 66420–66435, 2021, doi:
10.1109/TITS.2018.2889885. 10.1109/ACCESS.2021.3076679.
[11] K. Ginigeme and Z. Wang, ‘‘Distributed optimal vehicle-to-grid [30] Independent Electricity System Operator. IESO Market Manual Part
approaches with consideration of battery degradation cost under 5.5: Physical Markets Settlement Statements. Accessed: Jun. 2022.
real-time pricing,’’ IEEE Access, vol. 8, pp. 5225–5235, 2020, doi: [Online]. Available: [Link]
10.1109/ACCESS.2019.2963692. [Link]
[12] Y. Liu, P. Zhou, L. Yang, Y. Wu, Z. Xu, K. Liu, and X. Wang, [31] Q. Zhang, Y. Zhu, Z. Wang, Y. Su, and C. Li, ‘‘Reliability assessment of
‘‘Privacy-preserving context-based electric vehicle dispatching for energy distribution network and electric vehicle considering quasi-dynamic traffic
scheduling in microgrids: An online learning approach,’’ IEEE Trans. flow and vehicle-to-grid,’’ IEEE Access, vol. 7, pp. 131201–131213, 2019,
Emerg. Topics Comput. Intell., vol. 6, no. 3, pp. 462–478, Jun. 2022, doi: doi: 10.1109/ACCESS.2019.2940294.
10.1109/TETCI.2021.3085964. [32] The Atmospheric Fund. A Clearer View of Ontario’s Emissions.
[13] D. Wu, H. Zeng, C. Lu, and B. Boulet, ‘‘Two-stage energy management Accessed: Aug. 2022. [Online]. Available: [Link]
for office buildings with workplace EV charging and renewable energy,’’ uploads/2021/11/20211116_TAF_Emissions-[Link]
IEEE Trans. Transport. Electrific., vol. 3, no. 1, pp. 225–237, Mar. 2017, [33] J. Runyon. SDG&E Using V2G Technology for Emergency Load Shed
doi: 10.1109/TTE.2017.2659626. Events. Accessed: Aug. 2022. [Online]. Available: [Link]
[14] A. Ouammi, ‘‘Peak loads shaving in a team of cooperating smart [Link]/der-grid-edge/sdge-using-v2g-technology-for-emergency-load-
buildings powered solar PV-based microgrids,’’ IEEE Access, vol. 9, shed-events/
pp. 24629–24636, 2021, doi: 10.1109/ACCESS.2021.3057458. [34] V. S. Etchebehere and J. W. M. Lima, ‘‘Locational tariff structure for radial
[15] H. Dagdougui, A. Ouammi, and L. A. Dessaint, ‘‘Peak load reduction in network fixed costs in a DER context,’’ IEEE Access, vol. 10, pp. 597–607,
a smart building integrating microgrid and V2B-based demand response 2022, doi: 10.1109/ACCESS.2021.3137092.
scheme,’’ IEEE Syst. J., vol. 13, no. 3, pp. 3274–3282, Sep. 2019, doi:
10.1109/JSYST.2018.2880864.
[16] H. Shin and R. Baldick, ‘‘Plug-in electric vehicle to home (V2H) operation
under a grid outage,’’ IEEE Trans. Smart Grid, vol. 8, no. 4, pp. 2032–2041, SHIVAM SAXENA (Student Member, IEEE)
Jul. 2017, doi: 10.1109/TSG.2016.2603502. received the Ph.D. degree in electrical engineering
[17] S. Li, C. Gu, J. Li, H. Wang, and Q. Yang, ‘‘Boosting grid efficiency and
and computer science from the Lassonde School
resiliency by releasing V2G potentiality through a novel rolling prediction-
of Engineering, York University, Toronto, Canada.
decision framework and deep-LSTM algorithm,’’ IEEE Syst. J., vol. 15,
no. 2, pp. 2562–2570, Jun. 2021, doi: 10.1109/JSYST.2020.3001630.
His current research interests include distributed
[18] J. Hu, C. Ye, Y. Ding, J. Tang, and S. Liu, ‘‘A distributed MPC to exploit control of distributed energy resources and trans-
reactive power V2G for real-time voltage regulation in distribution net- active energy.
works,’’ IEEE Trans. Smart Grid, vol. 13, no. 1, pp. 576–588, Jan. 2022,
doi: 10.1109/TSG.2021.3109453.
[19] K. Kaur, N. Kumar, and M. Singh, ‘‘Coordinated power control of elec-
tric vehicles for grid frequency support: MILP-based hierarchical con-
trol design,’’ IEEE Trans. Smart Grid, vol. 10, no. 3, pp. 3364–3373,
May 2019, doi: 10.1109/TSG.2018.2825322. HANY E. Z. FARAG (Senior Member, IEEE) is
[20] M. S. H. Nizami, M. J. Hossain, and K. Mahmud, ‘‘A coordinated elec- currently an Associate Professor with the Depart-
tric vehicle management system for grid-support services in residential ment of Electrical Engineering and Computer
networks,’’ IEEE Syst. J., vol. 15, no. 2, pp. 2066–2077, Jun. 2021, doi: Science, Lassonde School of Engineering, York
10.1109/JSYST.2020.3006848. University. His current research interests include
[21] M. Aryanezhad, ‘‘Optimization of grid connected bidirectional V2G microgrids and transit electrification.
charger based on multi-objective algorithm,’’ in Proc. 8th Power Elec-
tron., Drive Syst. Technol. Conf. (PEDSTC), Feb. 2017, pp. 519–524, doi:
10.1109/PEDSTC.2017.7910381.
[22] S. Gao, H. Li, J. Jurasz, and R. Dai, ‘‘Optimal charging of electric vehicle
aggregations participating in energy and ancillary service markets,’’ IEEE
J. Emerg. Sel. Topics Ind. Electron., vol. 3, no. 2, pp. 270–278, Apr. 2022,
doi: 10.1109/JESTIE.2021.3102417.
[23] B. K. Sovacool, L. Noel, J. Axsen, and W. Kempton, ‘‘The neglected social LEIGH ST. HILAIRE (Member, IEEE) received
dimensions to a vehicle-to-grid (V2G) transition: A critical and systematic the bachelor’s degree in electrical and power engi-
review,’’ Environ. Res. Lett., vol. 13, no. 1, pp. 13001–13019, Jan. 2018, neering from Queen’s University, in 2008. His cur-
doi: 10.1088/1748-9326/aa9c6d. rent research interest includes the development of
[24] R. Ghotge, K. Nijssen, J. Annema, and Z. Lukszo, ‘‘Use before you choose: affordable net-zero housing.
What do EV drivers think about V2G after experiencing it?’’ Energies,
vol. 15, no. 13, pp. 4907–4928, Jul. 2022, doi: 10.3390/en15134907.
[25] C. Gschwendtner, S. R. Sinsel, and A. Stephan, ‘‘Vehicle-to-X (V2X)
implementation: An overview of predominate trial configurations and
technical, social and regulatory challenges,’’ Renew. Sustain. Energy Rev.,
vol. 145, pp. 110977–110994, Jul. 2021, doi: 10.1016/[Link].2021.110977.
[26] A. A. Mohamed, C. Sabillon, A. Golriz, and B. Venkatesh, ‘‘Value-stack
aggregator optimal planning considering disparate DERs technologies,’’
IET Gener., Transmiss. Distrib., vol. 15, no. 18, pp. 2632–2644, May 2021, AIDAN BROOKSON received the [Link].
doi: 10.1049/gtd2.12205. degree in mechanical engineering from Ryerson
[27] G. R. Parsons, M. K. Hidrue, W. Kempton, and M. P. Gardner, ‘‘Will- University. His current research interest includes
ingness to pay for vehicle-to-grid (V2G) electric vehicles and their optimization algorithms dedicated to decarboniz-
contract terms,’’ Energy Econ., vol. 42, pp. 313–324, Mar. 2014, doi: ing energy systems.
10.1016/[Link].2013.12.018.
[28] Y. Yu, Z. Cai, and Y. Huang, ‘‘Energy storage arbitrage in grid-
connected micro-grids under real-time market price uncertainty: A double-
Q learning approach,’’ IEEE Access, vol. 8, pp. 54456–54464, 2020, doi:
10.1109/ACCESS.2020.2981543.

VOLUME 11, 2023 17095