Emotional Factor Forecasting based on Driver Modelling in Electric

Vehicle Fleets

J. I. Guerrero

1 a

, M. C. Romero-Ternero

, E. Personal

, D. F. Larios

, J. A. Guerra

and C. León

Department of Electronic Technology, Universidad de Sevilla, C/ Virgen de África, Seville, Spain

Keywords: Emotional Computing, Driver Modelling, Electric Vehicle Fleet, Evolutionary Computation, Virtual Power

Plant, Smart Grid.

Abstract: Until recently, the automotive industry focus has been safety, comfort, and user experience. Now, the focus

is shifting towards human emotion for driver-car interactions, autonomy and sustainability; all of them are

increasing concerns in recent scientific literature. On the one hand, the growing role of emotion in automotive

driving is empowering human-centred design coupled with affective computing in driving context to improve

future automotive design. It is resulting in emotional analysis being present in automotive. This requires real-

time data processing that involves energy consumption in the vehicle. On the other hand, electric vehicle

fleets and smart grids are technologies that have provided new possibilities to reduce pollution and increase

energy efficiency looking for sustainability. This paper proposes the emotional factor forecasting according

to data gathered from electric vehicle fleet, based on the application of K-means algorithm. The results shows

that is possible to forecast the emotional status that takes negative effect in the driving. Additionally, the

Cronbach alpha variation analysis provides an interesting tool to select features from samples.

1 INTRODUCTION

For different reasons and purposes, the number of

studies related to include emotional analysis in cars is

growing in the scientific literature (Akamatsu et al.,

2013; Braun et al., 2019; Izquierdo-Reyes et al.,

2018; Khan & Lee, 2019; Nass et al., 2005; Schuller

et al., 2006). User-centred design coupled with

affective computing is resulting in emotional analysis

being present in cars.

In vehicles that consider the emotional indicators

of passengers, it is necessary to perform several

analyses (signal processing, feature extraction,

emotional classification and behaviour for reaction).

That processing requires energy consumption from

on-board computer. In the case of electric vehicles

(EV), it is interesting to forecast the consumption of

said processing in order to know how this can affect

https://orcid.org/0000-0003-3986-9267

https://orcid.org/0000-0001-6965-9485

https://orcid.org/0000-0003-3287-3300

https://orcid.org/0000-0002-4309-6028

https://orcid.org/0000-0002-7845-4446

https://orcid.org/0000-0002-0043-8104

to the autonomy of the vehicle and to the longest route

that can be made without recharging.

EVs represents a new research field in smart grid

(SG) ecosystems. Currently, the new generation of

EVs provides different technologies which can be

integrated in SGs. However, these new technologies

make difficult the distribution of grid management. In

particular, EVs and the infrastructure needed to

charge them have resulted in a great quantity of new

standards and technologies.

Currently, there are several research lines related

to EVs: fast charging networks, battery performance

modelling, parasitic energy consumption, EV

promotional policies, increasing the range of the

battery in EV, etc.; and other research lines related to

EV energy management: contract models for

consumption vehicle, market model to adopt EVs,

distributed energy resources management systems

Guerrero, J., Romero-Ternero, M., Personal, E., Larios, D., Guerra, J. and León, C.

Emotional Factor Forecasting based on Driver Modelling in Electric Vehicle Fleets.

DOI: 10.5220/0009561406030612

In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 1, pages 603-612

ISBN: 978-989-758-423-7

603

(DERMS), distributed energy resources (DER)

standards, faster charging technologies, demand

response management systems (DRMS), the role of

aggregators in V2G (vehicle-to-grid), energy

efficiency, customer support, driver support, etc.

Additionally, all these lines are influenced by the

current regulation and may different greatly between

countries (for example, the regulation between United

States (Lazar, s. f.) and Europe (CEER, 2019) is very

different regarding energy management).

The charging infrastructure affects SG on several

levels (Guerrero, Personal, García, et al., 2019;

Guerrero, Personal, Parejo, et al., 2019). These levels

concern the transmission, distribution, and retailer

levels. The main affected frameworks inside these

levels are: energy management (EM), distribution

management (DM), and demand response (DR). The

EM systems include several functions, one of which

is the control of energy flows. The charging of an EV

can be made at any point on the grid which has a

charging unit. If the system has information about the

expected use of the charging unit, the energy flow

will be easier to manage. The DM is related with

Distribution System Operators (DSO). Usually, the

charging infrastructure is overseen by the DSOs.

Thus, the DSOs must manage these facilities and

maintain information about them. Finally, the

demand response concerns retailers and DSOs, and

the main problem is demand curve flattening and

price management. Nevertheless, the new paradigm

proposed by standard organizations, including

National Institute of Standards and Technology

(NIST), International Electrotechnical Commission

(IEC), among others, related with V2G proposed that

EVs could charge or discharge batteries. Thus, the EV

is a power source in specific scenarios. In these cases,

the distributed resource management is affected by

the new V2G technologies as a distributed power

resource in low voltage without total availability, like

some renewable energy resources, for example, wind

and solar energy.

Our researching group has proposed a distributed

charging prioritization methodology based on the

concept of virtual power plant without considering

emotional factors consumption (Guerrero, Personal,

García, et al., 2019; Guerrero, Personal, Parejo, et al.,

2019). In these papers, we describe the Driver

Modelling module which is one of the elements of the

distributed charging prioritization methodology.

Additionally, we get advantage from the driver

pattern to stablish an emotional analysis based on the

deviation from the driver pattern.

Firstly, background of this study is presented.

Secondly, the methodology including emotional

factors consumption is described and finally some

conclusions and future work are outlined.

2 BACKGROUND

2.1 Electric Vehicle Fleets and Related

Technologies

The introduction of EVs provides several advantages,

but it is necessary to have additional energy sources

in order to include the associated infrastructure

(Meissner & Richter, 2003; Tie & Tan, 2013). The

new generation of EVs has several requirements not

only in power but also infrastructure (Francesco

Marra et al., 2011). SGs have provided a good

scenario to integrate EV and its charging

infrastructure.

Dielmann & Velden (2003) propose Virtual

Power Plant (VPPs) as a new solution for the

implementation of technologies related to SGs, and

several applications were developed to show the

advantages of VPPs. The FENIX European Project

(Kieny et al., 2009) delved into the concept of VPP

and considered two types of VPP: the commercial

VPP (CVPP), that tackles the aggregation of small

generating units with respect to market integration,

and the technical VPP (TVPP), that tackles

aggregation of these units with respect to services that

can be offered to the grid. Mashhour & Moghaddas-

Tafreshi (2009) described a general framework for

future VPP to control low and medium voltage for

DER management. You et al. (2009) presented a case

study which shows how a broker GVPP was

developed based on the selection of appropriate

functions. The EDISON Danish project (Binding

et al., 2010) described an ICT-based distributed

software integration based on VPPs and standards to

accommodate communication and optimize the

coordination of EV fleets. Jansen et al. (2010)

proposed an architecture for EV fleet coordination

based on V2G integrating VPP. Musio et al. (2010)

analysed the possibility of using EVs as an energy

storage system (V2G) within a VPP structure.

Skarvelis-Kazakos et al. (2010) considered the EV as

a mobile load and described a VPP containing

aggregated microgeneration sources and EV, but is

cantered around minimizing carbon emissions. Raab

et al. (2011) proposed and discussed three approaches

for grid integration of EVs through a VPP: control

structure, resource type, and aggregation. Sanduleac

et al. (2011) presented a solution for integrating EVs

in the SG through unbundled smart metering and VPP

technology dealing with multiple objectives. Marra et

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

604

al. (2012) addressed the design of an EV test bed

which served as a multifunctional grid-interactive EV

to test VPP or a generic EV coordinator with different

control strategies.

The common point of these references is the

utilization of the VPP concept in a simulation, but

they only simulate the VPP which aggregates the

information of EV. The present paper additionally

analyses the impact in VPP of higher levels, and how

the distribution of charging is made.

Additionally, some researchers have studied the

impact of HEV and plug-in HEV (PHEV) (He et al.,

2012). In this sense, decentralized algorithms for

coordinating the charging of multiple EVs have

gained importance in recent years. Mansour et al.

(2015) compared several approaches based on

centralized, decentralized, and hybrid algorithm, with

the latter showing better results. Hiermann et al.

(2016) introduced the electric fleet size and mix

vehicle routing problem with time windows and

recharging stations (E-FSMFTW) to model decisions

to be made with regards to fleet composition and

vehicle routes, including the choice of recharging

times and locations. Hu et al. (2016) presented a

review and classification of methods for smart

charging of EVs for fleet operators, providing three

control strategies and their commonly used

algorithms. Additionally, they studied service

relationships between fleet operators and four other

actors in SGs.

All these works did not consider behaviour or

emotional parameters to forecast charging

requirements in EV. In the next section we describe

how emotional factors are present in automotive

industry and how their impact has evolved.

2.2 Emotional Factors in Automotive

Over time, automotive industry has evolved by

changing the approach based on technological

developments and user needs. For highly automated

vehicles where the driver still has an active role and

control is shared between the automobile and the

driver, the role of human-automobile interaction is

highly significant (Weber, 2018).

Cooperation between car and driver needs that

interaction happens on an affective level to create a

successful control loop. To keep the human informed,

car must understand and respond to human behaviour

and emotions (Braun et al., 2019). Therefore, a high

level of understanding of drivers is required (Khan &

Lee, 2019).

For instance, Nass et al. (2005) studied whether

characteristics of a car voice can affect driver

performance and affect concluding that when user

emotion matched car voice emotion (happy/energetic

and upset/subdued), drivers had fewer accidents,

attended more to the road (actual and perceived), and

spoke more to the car. They also discussed

implications for car design and voice user interface

design. Schuller et al. (2006) introduced novel

concepts and results considering the estimation of a

driver’s emotion by focusing on acoustic information.

Izquierdo-Reyes et al. (2018) proposed a multiagent-

based framework called ADMAS (Advanced Driver

Monitoring for Assistance System). This system

considers the typical stages in affective computing,

including data acquisition (signals from wearable,

images from cameras, audio from microphones),

signal processing (computer vision, natural language

processing, audio mining) for feature extraction and

emotional classification using an emotional model

and machine learning to predict emotional

behaviours.

Shaikh & Krishnan (2012) proposed a framework

to combine empirical models describing human

behaviour with the environment and system models.

They analysed the design for safe vehicle-driver

interaction and showed a case study involving semi-

autonomous vehicles where the driver fatigue were

factors critical to a safe journey.

Videla & Kumar (2020) presented an approach to

detect person fatigue using image processing with

machine learning. In particular, they combined two

methods: face recognition with Histograms of

Oriented Gradients (HOG) and Support Vector

Machine (SVM) and off-the-shelf face detectors and

facial landmark detectors together with a novel eye

and mouth metric.

Silva & Analide (2019) considered that comfort

evaluation depends on environment attributes,

physical attributes and also emotion recognition.

They proposed a multiagent-based computational

sustainability platform which manages contexts

supported by principles of computational

sustainability and the assurance of sustainable

scenarios. They consider social indicators based on

mood analysis.

In all cases, artificial intelligence processing

applied in affective computing, above all regarding

machine learning techniques, requires substantial

energy consumption (Strubell et al., 2019). In the case

of electric vehicles this impacts in their autonomy.

Therefore, modelling driver behaviour and emotion is

useful to further refine the prediction of vehicle

power consumption.

Emotional Factor Forecasting based on Driver Modelling in Electric Vehicle Fleets

605

3 ARCHITECTURE VIEW

A solution for EV Fleet Management Platform based

on the concept of a VPP and using distributed

evolutionary computation algorithms to optimize the

prioritization of EV fleets at different levels of SG

ecosystems has been proposed in previous works. The

proposed architecture and methodology are described

in detail in (Guerrero, Personal, García, et al., 2019;

Guerrero, Personal, Parejo, et al., 2019).

Additionally, this reference treats only one of the

modules related to Charging Prioritization Module,

which is based on several Artificial Intelligence

Algorithms:

▪ Genetic algorithm (GA).

▪ Genetic algorithm with evolution control

(GAEC) based on fitness evolution curve.

▪ Swarm intelligence based on particle swarm

optimization (PSO).

The objective of the present paper is to describe

in detail the process of driver modelling, from the

acquisition to the modelling stage. Additionally, the

driver model is applied in a local application to

determine the alteration of driver pattern,

recommended different actions according to the

forecasting emotional status.

The viewpoint of the proposed solution treats

vehicles as a mobile load. In this manner, the system

must have data about these loads and the charging

prioritization. Thus, the system will have information

about the expected consumption or the expected

generation of the resource (in the case of a fault in the

grid), such as a battery.

The proposed system works as a service for large

companies with EV fleets. Knowledge about the state

and prioritization of vehicles and driver patterns may

minimize the impact of charging loads. These

services provide new tariffs for retailers and new

policies for energy price management.

The conceptual architecture of the proposed

solution is shown in

, where several VPPs are included. The

information is aggregated on the lower level. Then,

the aggregated information is sent by each lower VPP

to a higher level. In this manner, each VPP aggregates

the data and services from lower VPPs to higher

VPPs. Each level may have one or more VPPs

depending on the needs at each level and the power

grid.

Figure 1: Scalability properties and information flow

between different VPP layers.

The information representation at different levels was

based on an extension of the common information

model (CIM) from IEC 61870, 61968, 62325, and

eMIX (Energy Market Information Exchange). The

interface information is based on the component

interface specification (CIS) from the IEC and

OpenADR from OASIS (Open Association for

System and Information Standards). The information

representation and interface description are beyond

the scope of this paper.

Each higher VPP can perform evolutionary

algorithms to generate commands or instructions to

modify the queues from lower VPPs. Additionally,

lower VPPs can perform the same evolutionary

algorithms to request resources from other VPPs to

prioritize the charging of vehicles that cannot be

charged at their charging stations.

The artificial intelligence is based on data mining

algorithms or techniques. Each level runs the data

mining algorithms depending on the available

computational resources or option configured in the

corresponding VPP. The level at which the VPP is

performed determines the availability of services and

data. In this paper for the platform, four levels are

proposed:

▪ Smart business VPP (SBVPP). This is the

lowest level. At this level, all information

about vehicles, routes, and drivers from the

same company is available. Thus, the

charging prioritization of the charging

stations and driver patterns of the company is

treated at this level. The state of charge (SoC)

is also calculated at this level.

▪ Distribution VPP (DVPP). At this level,

information is aggregated from lower levels,

and information about retailers and the

presence of charging stations is stored. This

information is sent to higher levels, such as

an energy VPP (EVPP). Additionally, the

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

606

restrictions from an EVPP to the

corresponding retailer and SBVPP are

addressed at this level.

▪ Retailer VPP (RVPP). At this level, a retailer

needs to know when vehicles require

charging at any point outside of the company

points. The retailer can use this information

to offer different tariffs to a company.

▪ Energy VPP (EVPP). In this paper, the

vehicles represent mobile loads. Thus, if an

energy management system has information

about the expected charging stations, it may

take advantage of this information to improve

the load flow forecasting algorithms.

The prioritization process is performed in several

stages to aggregate information for the upper layers

and to control lower layers.

The lowest levels implement functions that are

related to consumers. The medium levels implement

functions that are related to energy distribution and

commercialization. The higher levels implement

functions to guarantee the quality and continuity of a

power supply. This architecture is highly scalable,

increasing the interoperability between different

enterprises and integration of heterogeneous

ecosystems. The VPPs include data mining

algorithms and some capabilities in the corresponding

driver modelling modules which makes the driver

pattern modelling quicker and easier:

The data mining algorithm in an SBVPP. This

algorithm sorts the vehicles with their drivers

according to the SoC and expected route. If the

algorithm cannot model any driver, the algorithm

classifies the driver as an external model and sends

the request to higher VPPs. The SBVPP can receive

commands and warnings from the DVPP and RVPP,

and it downloads general driver patterns. The higher

VPP commands and warnings are considered as

external restrictions. The external driver patterns are

considered as general models with low priority level,

and they will be replaced by the models generated in

the first route. Additionally, the RVPP commands and

warnings can take effect over different elements of

customer power facilities when the customer that

implements a SBVPP has contracted additional

services from a retailer to manage the customer power

facilities.

The data mining in a DVPP. The DVPP gathers

all requests from all SBVPPs. In the prioritization

module, this information is employed in an

evolutionary algorithm to prioritize charging in

available charging stations. In case of driver

modelling module, the data mining algorithm takes

advantage form different modules generated in the

SBVPP, providing a generalized classification of the

different models or patterns. The generated models

are the basis for the driver models in the SBVPP level.

The EVPP does not gather any information from

driver modelling, but this level takes an important

role in the charging prioritization.

The RVPP gathers all information about vehicles

that may have contractual relationships with a

retailer. The retailer can use this information to offer

new services to clients. If any problems arise in the

client contract, the retailer can send a command or

alarm to change the prioritization for one or more

vehicles and/or charging stations. In case of driver

models, the RVPP gather information about the best

driver pattern (in terms of energy efficiency), and the

RVPP could offer new services or advantages related

to the correspondence with the driver pattern.

All information about driving is gathered from

vehicle and transmitted to the SBVPP when the driver

cellular connects to the acquisition system. The

information about the current route and data from the

vehicle is stored in the Driver Modelling. This

information is used to update the driver model or

pattern.

Any algorithm for the SBVPP and DVPP is

possible because the algorithm works independently

of other layers. Thus, several algorithms were tested

in this paper, and a final configuration is proposed

based on the results of the tests. However, the

algorithms can be configured according to the

resources of each level.

3.1 The Electric Vehicle Fleet

Management VPP

The Electric Vehicle Fleet Management VPP or Node

(EVFMN) is the generic system implemented in each VPP.

The architecture of Electric Vehicle Fleet Management

Platform (EVFMP) is shown in Figure 1, and it is formed

by the replication of EVFMN between different VPPs,

enabling or disabling certain functionalities according to

the level of VPP. The EVFMN is shown in Figure 2. Each

module has specific functions:

▪ Asset Management System. The asset

management system is based on the predictive

maintenance of vehicles and charging stations.

▪ Driver Modelling. This module executes a

modelling process of driver behaviour. This

module provides a driver pattern which is used to

schedule the routes and, in this case, to

forecasting the driver’s emotional context.

▪ Energy Efficiency. This module applies different

techniques to optimize the energy consumption

Emotional Factor Forecasting based on Driver Modelling in Electric Vehicle Fleets

607

Figure 2: Modules of distributed evolutionary prioritization

framework.

and reduce the maintenance periods and economic

impact.

▪ Real-Time Route Scheduling. This module

manages all information about vehicles, routes,

drivers, and external conditions to establish

better prioritization in each charging station.

▪ Information Management. This module manages

all information of this VPP for reporting and

visualization.

▪ Prioritization Algorithm. The prioritization

algorithm in this layer is based on swarm

intelligence.

▪ External Coordination. This module sends

information to higher layers and gathers

information about external requirements or

vehicles to charge.

The external coordination is provided by the

interoperability with higher VPP layers.

Some modules, such as external coordination,

prioritization algorithm, and the SoC module, are

available for all VPPs. The other modules depend on

the available information in the VPPs. For example,

the SBVPP has all information about the EV fleet;

however, the SBVPP may have additional services of

energy efficiency if it shares the information with the

RVPP (in this case, the RVPP would use the energy

efficiency module).

In case of Driver Modelling, it is not included in

the EVPP level, and it optionally could be included in

RVPP, depending on the services provided by this

level.

3.2 Information Acquisition from

Electric Vehicle

The information is gathered from ODB-II system

(Road vehicles—Diagnostic systems—Part 2: CARB

requirements for interchange of digital information,

1994), a standardized CAN-bus based protocol,

designed for cars monitoring. This bus was

introduced in 1995 in North America, being

mandatory in all cars since 2008 (Taha & Nasser,

2015). Similar situation happens in Europe, being

mandatory in all gasoline vehicles since 2001 and

diesel since 2003.

Therefore, according to different regulations all

modern relies on embedded computers, called engine

control units (ECUs) (Moore et al., 2017), designed

to control different subsystem of the vehicle as motor

control, lights, braking subsystem, etc., most of them

using standardized messages.

As appear in the literature, this information can be

used to estimate vehicle speed (Bagheri et al., 2018),

or modelling the behaviour of the driver (Wang et al.,

2018), as we need in our proposal.

In this case, we use an ODB-II to Bluetooth

interface that sends the information to a mobile

application that executes the proposed algorithm.

3.3 Driver Patterns

Driver behaviour is stored in driver patterns. The

driver pattern is a model that takes effect over the

consumption of a vehicle in route scheduling. The

driver pattern affects the calculated SoC for each

section of a route; it depends on the terrain topology

and traffic data. Driver behaviour is calculated

according to the historical data of a driver. If

historical information about a driver is not available,

this pattern will be calculated only with the emotional

information.

The driver pattern consists of the deviation from

the original predicted SoC. This pattern considers

information about traffic, weather and previous

emotional behaviour to explain the variation from the

original predicted SoC.

Although a default driver pattern can be defined,

information about driver patterns is currently

unavailable. A default “average” driver pattern can be

created when a system has adequate information.

Currently, this pattern does not include the utility

factor [39] because the EV fleets are treated as mobile

loads and they do not include PHEVs.

The Driver Modelling process is based on Generic

Rule Induction (GRI), Support Vector Machine

(SVM) and K-Means clustering algorithm involving

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

608

the Cronbach alpha variation analysis, which uses the

information provided by acquisition system:

accelerator and brake usage, average speed, variation

of revolution per minute (rpm), usage of HVAC in the

vehicle, brake energy restoration, etc. All this

information provides a classification of driver pattern,

which is translated into a set of typical values for

different parameters in a different confidence ranges

and correlation variation.

3.4 Real-time Route Scheduling

This module controls several conditions that can

modify the current prioritization charging queues.

This module notifies any change in the following

conditions:

▪ Driver and EV availability.

▪ Route modifications.

▪ Traffic and roadwork.

▪ Weather conditions.

▪ Charging station availability.

3.5 SoC Module: Estimation of EV

Consumption

The proposed solution is based on the instantaneous

SoC value of each EV. These algorithms require an

estimation of some consumption according to its

planned route and alternative routes to reach different

recharging spots. This consumption estimation is

supported by a route planning tool. However, these

estimations are not trivial and are related to the

distance or time of the trip (Shankar & Marco, 2013;

De Cauwer et al., 2015). Other factors (e.g., road

(Park et al., 2009) and vehicle characteristics, traffic

(Boriboonsomsin et al., 2012), driving style

(Bingham et al., 2012), and weather conditions) are

essential for this estimation.

4 DATA MINING ALGORITHMS

The data mining algorithms are based on the

combination of three algorithm or techniques:

▪ Cronbach alpha variation analysis. This

technique aids to classify the different

parameters according to their variance

related to other parameters, providing a map

of the importance of different parameters in

the driver pattern. The Cronbach alpha

calculated for all parameters provided a

general number, which describe the

correlative variance between parameters.

Additionally, the Cronbach alpha relative to

each parameter p is calculated, providing a

new value for Cronbach alpha

corresponding to the new correlative

variance, without the influence of the

corresponding parameter p. If the value is

greater than the general Cronbach alpha, the

parameter p has a low level of relation with

the driver pattern, and the variation of this

parameter could provide erroneous patterns.

If the value is lesser than the general

Cronbach alpha, the parameter p has a high

level of relation with the driver pattern, and

the parameter is probably affected by

emotional behaviour.

▪ K-means. Classifies the values according to

the Cronbach Alpha stablishing

classifications for each range of each

parameter. Additionally, the algorithm is

used to classify the different results of

patterns, making groups according to the

common characteristics of the driver

behaviour. The method supposed that the

driver drives with a pattern and this pattern

is related to emotional context. Thus, the

different groups describe different

emotional status. In this case, it is not

important to reveal the emotion. Thus,

according to the distribution the emotions

are named emotion1, emotion2, etc.

▪ Support Vector Machine. Classifies the

driver pattern according to the other driver

patterns.

This paper is centred in the results of K-mean

algorithm to get different clusters which represents a

classification of patterns based on emotions.

5 EXPERIMENTAL RESULTS

5.1 Sample Description

The sample provided comes from a real sample

extracted from vehicles. However, this information

was not extracted in real time, it was extracted by

using a device to gather periodically information from

vehicles, based on CAN-bus (Controller Area

Network), specified by different standards (J2411,

J2284, J1939, ISO 11898, etc.). The information is

complemented by information of routing

management and scheduling,

The extracted information comprises a sample

with 2711 different routes. Each route is done by 32

different drivers during three months, around the

same area.

Emotional Factor Forecasting based on Driver Modelling in Electric Vehicle Fleets

609

The sample contains the following information:

weather information, driver code, route code, route

distance, number of stops in the route, route

information, traffic information, doors, lights, brake,

accelerator, gearbox, speed, voltage, distance

travelled, current, and state of charge of the batteries.

Additionally, there are some claims notified by

clients about the drivers. This information is

modelled by a parameter with the number of claims

by route.

From Cronbach alpha variation analysis, two

parameters are associated to each variable: the

general Cronbach alpha, the correlation coefficient,

and the Cronbach alpha if the corresponding

parameter is removed from sample. These parameters

are used to check the importance of parameters in the

sample compared with the K-means results.

5.2 K-mean Results

The application of K-mean algorithm in the data

provided by an electric vehicle fleet in distribution

logistics, provided a basic classification of different

patterns to drive, which take effect in the efficiency

of driving.

The different clusters are corresponding to

different driving types (figure 3). For example: the

cluster-5 correspond to the emotion5, this emotion is

like the great bag in which all the cases that are not

possible to classify, including the 30% of claims (9

claims). However, the cluster-4 (emotion4) has 14,9

% of routes, and groups the drivers who has a very

aggressive driving (this information is extracted from

parameters of speed, accelerator, gearbox, and state

of charge), including the 70% of claims (21 claims).

Thus, using the information provided by vehicles and

making this classification, it is possible to provide a

forecasting about the influence of emotion in the

vehicle driving, and could be notified to the system,

in order to maintain a good level of efficiency in the

consumption of vehicle. Cluster-2 and cluster-3

corresponding to drivers with careless driving pattern,

according to the information from parameters. The

cluster-1 groups all the patterns which provokes a

high efficiency driving, decreasing the consumption

and an understandable time invested in the route.

The size of smallest cluster is cluster-3 with 0,6%

or 16 cases. The size of biggest cluster is cluster-5

with 53,9% or 1461 cases.

Figure 3: Sizes of Clusters obtained from the application of

K-mean algorithm.

6 CONCLUSIONS AND FUTURE

RESEARCH

The proposed platform can integrate information

from electric vehicles to be considered as part of the

electric vehicle fleet management platform to

integrate the electric vehicle fleets in smart grids as

mobile loads.

On one hand, the role of emotion in automotive

driving is increasingly present, empowering human-

centred design coupled with affective computing in

driving context to improve future automotive design.

The driver emotional status influence is modelling by

the deviation of the driver pattern based on a Generic

Rule Induction, Support Vector Machine and K-

Means clustering algorithm involving the Cronbach

alpha variation analysis, which provides a light-

weight model to perform in low feature devices.

On the other hand, electric vehicle fleets and

smart grids are technologies that have provided new

possibilities to reduce pollution and increase energy

efficiency looking for sustainability. The inclusion of

driving data can improve the routing and charging

prioritization forecasting, providing additional

services to the different actors in the energy market,

and other advantages for the better stability of the

power grid.

The future works will be centred on provide more

information about driver, including some devices to

provide more information about emotional status of

driver, based on biometric factors or emotional

estimation by means of face image analysis. This

information provides more possibilities to analyse the

results presented in the present paper.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

610

REFERENCES

Akamatsu, M., Green, P., & Bengler, K. (2013).

Automotive Technology and Human Factors Research:

Past, Present, and Future. International Journal of

Vehicular Technology, 2013, 27.

Bagheri, S. M., Assemi, B., Mesbah, M., & Hickman, M.

(2018). Cost-effective ubiquitous method for motor

vehicle speed estimation using smartphones. IET

Wireless Sensor Systems, 8(6), 340-349.

https://doi.org/10.1049/iet-wss.2018.5023

Binding, C., Gantenbein, D., Jansen, B., Sundström, O.,

Andersen, P., Marra, F., Poulsen, B., & Træholt, C.

(2010). Electric vehicle fleet integration in the Danish

EDISON project—A virtual power plant on the island

of Bornholm. 1-8.

https://doi.org/10.1109/PES.2010.5589605

Bingham, C., Walsh, C., & Carroll, S. (2012). Impact of

driving characteristics on electric vehicle energy

consumption and range. IET Intelligent Transport

Systems, 6(1), 29-35. https://doi.org/10.1049/iet-

its.2010.0137

Boriboonsomsin, K., Barth, M. J., Zhu, W., & Vu, A.

(2012). Eco-Routing Navigation System Based on

Multisource Historical and Real-Time Traffic

Information. IEEE Transactions on Intelligent

Transportation Systems, 13(4), 1694-1704.

https://doi.org/10.1109/tits.2012.2204051

Braun, M., Schubert, J., Pfleging, B., & Alt, F. (2019).

Improving Driver Emotions with Affective Strategies.

Multimodal Technologies and Interaction, 3(1), 21.

https://doi.org/10.3390/mti3010021

CEER. (2019). Report on Regulatory Frameworks for

European Energy Networks, CEER (N.

C18-IRB-38-

03). https://www.ceer.eu/documents/104400/-/-

/9665e39a-3d8b-25dd-7545-09a247f9c2ff

De Cauwer, C., Van Mierlo, J., & Coosemans, T. (2015).

Energy Consumption Prediction for Electric Vehicles

Based on Real-World Data. Energies, 8(8), 8573-8593.

https://doi.org/10.3390/en8088573

Dielmann, K., & Velden, A. (2003). Virtual power plants

(VPP)—A new perspective for energy generation? 18-

20. https://doi.org/10.1109/SPCMTT.2003.1438108

Guerrero, J. I., Personal, E., García, A., Parejo, A., Pérez,

F., & León, C. (2019). Distributed Charging

Prioritization Methodology Based on Evolutionary

Computation and Virtual Power Plants to Integrate

Electric Vehicle Fleets on Smart Grids. Energies,

12(12), 2402. https://doi.org/10.3390/en12122402

Guerrero, J. I., Personal, E., Parejo, A., García, S., García,

A., & León, C. (2019). Forecasting Recharging

Demand to Integrate Electric Vehicle Fleets in Smart

Grids. Advanced Communication and Control Methods

for Future Smartgrids.

https://doi.org/10.5772/intechopen.88488

He, Y., Chowdhury, M., Ma, Y., & Pisu, P. (2012). Merging

mobility and energy vision with hybrid electric vehicles

and vehicle infrastructure integration.

https://doi.org/10.1016/j.enpol.2011.11.021

Hiermann, G., Puchinger, J., Ropke, S., & Hartl, R. F.

(2016). The Electric Fleet Size and Mix Vehicle

Routing Problem with Time Windows and Recharging

Stations. European Journal of Operational Research,

252(3), 995-1018.

https://doi.org/10.1016/j.ejor.2016.01.038

Hu, J., Morais, H., Sousa, T., & Lind, M. (2016). Electric

vehicle fleet management in smart grids: A review of

services, optimization and control aspects. Renewable

and Sustainable Energy Reviews, 56, 1207-1226.

https://doi.org/10.1016/j.rser.2015.12.014

Road vehicles—Diagnostic systems—Part 2: CARB

requirements for interchange of digital information,

ISO 9141-2:1994 (1994).

Izquierdo-Reyes, J., Ramirez-Mendoza, R. A., Bustamante-

Bello, M. R., Navarro-Tuch, S., & Avila-Vazquez, R.

(2018). Advanced driver monitoring for assistance

system (ADMAS). International Journal on Interactive

Design and Manufacturing (IJIDeM), 12(1), 187-197.

https://doi.org/10.1007/s12008-016-0349-9

Jansen, B., Binding, C., Sundström, O., & Gantenbein, D.

(2010). Architecture and Communication of an Electric

Vehicle Virtual Power Plant. 2010 First IEEE

International Conference on Smart Grid

Communications, 149-154.

https://doi.org/10.1109/SMARTGRID.2010.5622033

Khan, M. Q., & Lee, S. (2019). A Comprehensive Survey

of Driving Monitoring and Assistance Systems. Sensors

(Basel, Switzerland), 19(11).

https://doi.org/10.3390/s19112574

Kieny, C., Berseneff, B., Hadjsaid, N., Besanger, Y., &

Maire, J. (2009). On the Concept and Interest of Virtual

Power Plant: Some Results from the European Project

FENIX. 1-6.

https://doi.org/10.1109/PES.2009.5275526

Lazar, J. (s. f.). Electricity Regulation In the US: A Guide.

228.

Mansour, S., Harrabi, I., Maier, M., & Joós, G. (2015). Co-

simulation study of performance trade-offs between

centralised, distributed, and hybrid adaptive PEV

charging algorithms. Computer Networks: The

International Journal of Computer and

Telecommunications Networking, 93(P1), 153–165.

https://doi.org/10.1016/j.comnet.2015.08.036

Marra, F., Sacchetti, D., Pedersen, A. B., Andersen, P. B.,

Træholt, C., & Larsen, E. (2012). Implementation of an

Electric Vehicle test bed controlled by a Virtual Power

Plant for contributing to regulating power reserves.

2012 IEEE Power and Energy Society General

Meeting, 1-7.

https://doi.org/10.1109/PESGM.2012.6345269

Marra, Francesco, Sacchetti, D., Træholt, C., & Larsen, E.

(2011). Electric vehicle requirements for operation in

smart grids. 1-7.

https://doi.org/10.1109/ISGTEurope.2011.6162648

Mashhour, E., & Moghaddas-Tafreshi, S. M. (2009). The

opportunities for future virtual power plant in the power

market, a view point. 2009 International Conference on

Clean Electrical Power, 448-452.

https://doi.org/10.1109/ICCEP.2009.5212014

Emotional Factor Forecasting based on Driver Modelling in Electric Vehicle Fleets

611

Meissner, E., & Richter, G. (2003). Battery Monitoring and

Electrical Energy Management: Precondition for future

vehicle electric power systems. Journal of Power

Sources, 116(1), 79-98. https://doi.org/10.1016/S0378-

7753(02)00713-9

Moore, M. R., Bridges, R. A., Combs, F. L., Starr, M. S., &

Prowell, S. J. (2017). Modeling inter-signal arrival

times for accurate detection of CAN bus signal injection

attacks: A data-driven approach to in-vehicle intrusion

detection. Proceedings of the 12th Annual Conference

on Cyber and Information Security Research, 1–4.

https://doi.org/10.1145/3064814.3064816

Musio, M., Lombardi, P., & Damiano, A. (2010). Vehicles

to grid (V2G) concept applied to a Virtual Power Plant

structure. The XIX International Conference on

Electrical Machines - ICEM 2010, 1-6.

https://doi.org/10.1109/ICELMACH.2010.5608261

Nass, C., Jonsson, I.-M., Harris, H., Reaves, B., Endo, J.,

Brave, S., & Takayama, L. (2005). Improving

Automotive Safety by Pairing Driver Emotion and Car

Voice Emotion. CHI ’05 Extended Abstracts on Human

Factors in Computing Systems, 1973–1976.

https://doi.org/10.1145/1056808.1057070

Park, J., Chen, Z., Kiliaris, L., Kuang, M. L., Masrur, M.

A., Phillips, A. M., & Murphey, Y. L. (2009).

Intelligent Vehicle Power Control Based on Machine

Learning of Optimal Control Parameters and Prediction

of Road Type and Traffic Congestion. IEEE

Transactions on Vehicular Technology, 58(9), 4741-

4756. https://doi.org/10.1109/tvt.2009.2027710

Raab, A. F., Ferdowsi, M., Karfopoulos, E., Unda, I. G.,

Skarvelis-Kazakos, S., Papadopoulos, P., Abbasi, E.,

Cipcigan, L. M., Jenkins, N., Hatziargyriou, N., &

Strunz, K. (2011). Virtual Power Plant Control

concepts with Electric Vehicles. 2011 16th

International Conference on Intelligent System

Applications to Power Systems, 1-6.

https://doi.org/10.1109/ISAP.2011.6082214

Sanduleac, M., Eremia, M., Toma, L., & Borza, P. (2011).

Integrating the Electrical Vehicles in the Smart Grid

through unbundled Smart Metering and multi-objective

Virtual Power Plants. 2011 2nd IEEE PES

International Conference and Exhibition on Innovative

Smart Grid Technologies, 1-8.

https://doi.org/10.1109/ISGTEurope.2011.6162786

Schuller, B., Lang, M., & Rigoll, G. (2006). Recognition of

Spontaneous Emotions by Speech within Automotive

Environment. 57-58.

Shaikh, S., & Krishnan, P. (2012). A Framework for

Analysing Driver Interactions with Semi-Autonomous

Vehicles. Electronic Proceedings in Theoretical

Computer Science, 105, 85-99.

https://doi.org/10.4204/EPTCS.105.7

Shankar, R., & Marco, J. (2013). Method for estimating the

energy consumption of electric vehicles and plug-in

hybrid electric vehicles under real-world driving

conditions. IET Intelligent Transport Systems, 7(1),

138-150. https://doi.org/10.1049/iet-its.2012.0114

Silva, F., & Analide, C. (2019). Computational

sustainability and the PHESS platform: Using affective

computing as social indicators. Future Generation

Computer Systems, 92, 329-341.

https://doi.org/10.1016/j.future.2018.10.006

Skarvelis-Kazakos, S., Papadopoulos, P., Grau, I., Gerber,

A., Cipcigan, L. M., Jenkins, N., & Carradore, L.

(2010). Carbon optimized Virtual Power Plant with

Electric Vehicles. 45th International Universities

Power Engineering Conference UPEC2010, 1-6.

Strubell, E., Ganesh, A., & McCallum, A. (2019). Energy

and Policy Considerations for Deep Learning in NLP.

Proceedings of the 57th Annual Meeting of the

Association for Computational Linguistics, 3645–3650.

https://doi.org/10.18653/v1/P19-1355

Taha, A.-E. M., & Nasser, N. (2015). Utilizing CAN-Bus

and smartphones to enforce safe and responsible

driving. 2015 IEEE Symposium on Computers and

Communication (ISCC), 111-115.

https://doi.org/10.1109/ISCC.2015.7405502

Tie, S. F., & Tan, C. W. (2013). A review of energy sources

and energy management system in electric vehicles.

Renewable and Sustainable Energy Reviews, 20, 82-

102. https://doi.org/10.1016/j.rser.2012.11.077

Videla, L. S., & Kumar, M. A. (2020). Fatigue Monitoring

for Drivers in Advanced Driver-Assistance System.

Examining Fractal Image Processing and Analysis,

170-187. https://doi.org/10.4018/978-1-7998-0066-

8.ch008

Wang, W., Xi, J., & Zhao, D. (2018). Learning and

Inferring a Driver’s Braking Action in Car-Following

Scenarios. IEEE Transactions on Vehicular

Technology, 67(5), 3887-3899.

https://doi.org/10.1109/TVT.2018.2793889

Weber, M. (2018). Automotive emotions: A human-centred

approach towards the measurement and understanding

of drivers’ emotions and their triggers [Thesis, Brunel

University London].

http://bura.brunel.ac.uk/handle/2438/16647

You, S., Træholt, C., & Poulsen, B. (2009). Generic Virtual

Power Plants: Management of distributed energy

resources under liberalized electricity market. 1-6.

https://doi.org/10.1049/cp.2009.1764

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

612