Understanding Behavioural Conflict between the Drivers and

Adaptive Cruise Control (ACC) System in Cut-in Scenario

Jing Gong

, Fang You

, Jian-min Wang

and Xiao-long Zhang

Car Interaction Design Lab, Tongji University, Caoan Highway, Shanghai, China

College of Information Science and Technology, Pennsylvania State University, University Park, PA, U.S.A.

Keywords: Adaptive Cruise Control (ACC), Human-Machine Interface Design, Human-Machine Cooperation,

Human-Machine Conflict.

Abstract: In the cut-in scenario of the ACC system, there is often a lack of harmony between people and cars due to the

limitations of sensors and control strategies. Finding and solving the conflict between the driver and the

machine is essential to achieve harmonious Human-Machine Cooperation. This research is to understand the

conflict between the driver and ACC system in the cut-in scenario based on the previous work of driver trust

experiment. The research selected eight drivers for in-depth interview, and the results showed that the biggest

conflict between the driver and ACC was that the driver's cognitive and behavioural patterns were

significantly different from the ACC system. It is mainly reflected on three aspects: the different definition of

the cut-in scenario, the risk perception and the stress of the impending danger, and the perceptual process of

cut-in scenario. In order to reduce human-machine conflict, the research proposed three design strategies: (1)

Redefine the cut-in scenario based on the driver's cognition. (2) Keep the ACC human-machine interface

consistent with the driver's psychological perception. (3) Help drivers cope with dangerous scenario with three

levels of warning signals: guidance information, warning information and takeover information.

1 INTRODUCTION

The Advanced Driving Assistant System (ADAS) can

help drivers reduce burdens and improve safety,

comfort and convenience in driving, and has been

widely used in recent years. With the development of

sensor technology and data acquisition and

processing technology, sensory deficits of the driver

are compensated by technical sensors; for example,

the limited vision at night of drivers is enhanced by

night vision systems. These domains are commonly

referred to Human-Machine Cooperation (HMC) and

Cooperative Systems (COS) (Bengler K et al., 2012).

Hoc (Hoc, 2001) has detailed a precise definition of

cooperation that is consistent with most of the

literature:

Two agents are in a cooperative situation if they

meet two minimal conditions. (1) Each one strives

towards goals and can interfere with the other on

goals, resources, procedures, etc. (2) Each one tries to

manage the interference to facilitate the individual

activities and/or the common task when it exists. The

symmetric nature of this definition can be only partly

satisfied (Hoc, 2001).

The Adaptive Cruise Control (ACC) is a

comfortable, intelligent driving assist that can

maintain a steady state of driving depending on the

driver's pre-set speed and pre-set distance. But in

some scenarios, there are still limitations. As shown

in Fig.1, the cut-in vehicle changes lane and becomes

the new front vehicle for the ACC system, but the

ACC is not always quick to identify and lock the front

Figure 1: The response of ACC in cut-in scenario.

Gong, J., You, F., Wang, J. and Zhang, X.

Understanding Behavioural Conﬂict between the Drivers and Adaptive Cruise Control (ACC) System in Cut-in Scenario.

DOI: 10.5220/0008053600970103

In Proceedings of the 3rd International Conference on Computer-Human Interaction Research and Applications (CHIRA 2019), pages 97-103

ISBN: 978-989-758-376-6

vehicle. If the speed at which the ACC system is set

faster than the current speed of the vehicle, the

vehicle will continue to accelerate until one of the

following events occurs: the system reaches its set

speed; the system recognizes and catches the new

guided vehicle; the driver takes over the ACC system

(Larsson et al., 2014).

The driver takes over ACC without being

prompted, which means that the disruption of human-

machine cooperation of controling the vehicle. The

study by Frank et al., (2012) suggests that the

dynamic balance of the human-machine system must

have a cycle of perception, decision-making, and

action (Fig.2). In the perceptual stage, the difference

between the driver and the human-machine system is

the greatest. Firstly, the two are different to the

perception of the environment and the judgement of

the danger. Secondly, the cut-in scenario is a short

dynamic process, which may increase the cognitive

load of drivers, and requires the driver to have

sufficient situational awareness and reaction ability.

Thirdly, the takeover itself is a behaviour that requires

high perception load, which leads to a longer reaction

time (Tsang-Wei Lin, 2009). All of these has

implications for drivers taking over ACC.

Figure 2: The response of ACC in cut-in scenario.

In previous studies. Pengjun et al., (2005) showed

that the degree to which drivers take over ACC

manually depends on a number of factors, and an

important one is the difference between the ACC's

performance and the driver's expectations. This is

consistent with Wouter (Wouter et al., 2017) and

Bobbie (Bobbie et al., 2015). In Brian's research

(Brian et al., 2012), it is pointed out that how drivers

make decisions is not based on external

circumstances, but based on perceived mental

representations. In the cut-in scenario, the driver's

mental representation is consistent with the ACC's

control strategy, which directly leads to the driver

taking over the vehicle to avoid risk. Because the

psychological model of different drivers is different,

their mental representations will be changed

according to the degree of familiarity and degree of

trust of the ACC system. It is almost impossible to get

the ACC system's control strategy exactly the same as

the driver's mental model, but it can reduce the

difference between the perceptual stage in the human-

machine interface: not only should the driver be

notified of the behaviour and state of the ACC

system, but also the driver should be guided by the

driver's perception of the cut-in scenario.

In recent years, scholars have come up with new

ideas about how to design Intelligent Manufacturing

Systems (IMS), replacing techno-centered design

with human-centered design. Some scholars have

worked on Human-Machine Cooperation, Levels of

Automation and Situation Awareness, and Human-

Automation Symbiosis (May et al., 2014; Millot,

2014; Romero et al., 2015.) (Pacaux-Lemoine et al.,

2017). In order to improve the efficiency of human-

computer interaction, the study of Bjørn Solvang et

al., (2012) discusses the cooperative mode and

interaction between human operators and machines,

and presents an open control system for new and old

equipments. The study by Oborski (Oborski et al.,

2004) suggests that the whole system performance

depends on human decisions, and the significant

stress should be put on the problem of human-

machine and human-computer systems co-operation.

In Pacaux’ study (Pacaux et al., 2017) proposed a

principle to retain humans in the process control loop

with different levels of involvement identified by the

levels of automation.

This research is also based on the idea of human-

centered design, studying how people and machines

can collaborate better and improve people's

productivity.

2 RESEARCH METHOD

2.1 ACC Cut-in Scenario Analysis

This study discusses the human-machine relationship

and driving condition of the driver and ACC system

in the cut-in scenario. It is pointed out that the human-

machine collaborative control is the ideal state of

human-machine interaction, which provides the basic

theoretical guidance for in-depth interview and

human-machine interface design.

The state of human-machine driving can be

divided into four types according to the role of driver

and machine in driving a vehicle: both of them leave

it alone; the vehicle is driven by the intelligent driving

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

Figure 3: Human-Machine cooperative driving status of ACC in cut-in scenario.

system (ACC); driver controls the vehicle; driver and

intelligent driving system (ACC) work together to

control vehicles through human-machine interface.

As shown in Fig.3, in the ACC cut-in scenario,

these four conditions are:

(1) The machine is left unattended: because of the

shortcomings of the ACC sensor, it was unable to

identify the side vehicles. And because of the driver's

trust in ACC, the driver may not have noticed the side

vehicle during the side car's cutting-in.

(2) Intelligent driving system (ACC) controls the

vehicle: the vehicle senses the cut-in vehicle and

processes it before the driver intervenes. The premise

of this is that the ACC is fully trusted, but this leads to

two results: the ACC's steady and safe completion of

the vehicle task; and the ACC's delay in the driver's

expectation, which reduces the driver's trust in the

ACC system.

(3) The driver controls the vehicle: the vehicle senses

the cut-in vehicle, but the brake time is later than the

driver's expectation, or the two vehicles’ distance is

less than the driver's safety expectation, and the driver

takes over vehicle initiatively. Another is when the

car and the side car driving normally. However, people

have the ability to predict and feel that the side car want

to cut in, then he takes over in advance when the side

car gets closer and closer to the side lane line.

(4) Human-machine cooperative control: the vehicle

will perceive information such as the current state and

behaviour of ACC, current environment information,

etc., through the interactive interface to the driver.

ACC system can actively decelerate, and human

interface can warn the driver in advance, so that the

driver can take over in time.

2.2 Experimental Design and In-depth

Interview

The interview was based on a previous work that

researches the driver's trust in the ACC in cut-in

scenario. The experiment uses v-box to capture the

actual driving video (the dashboard interface and the

external environment, both in sync) when the Volvo

XC90 turns on the ACC and cuts in. As shown in fig.4,

the experiment simulated real-life road scenes on a

driving simulator, and placed the dashboard video on

the iPad in front of the steering wheel to simulate the

real dashboard. The experiment has set up six

conditions, respectively, the speed of the car was

V=30km/h, 50km/h, 60km/h, and the distance between

the two vehicles was THW=0.7s, 1.2s. By evaluating

drivers ' trust in ACC to understand the relationship

between drivers and ACC, the results showed that the

speed and THW values had a definite effect on the

level of trust that the driver had when he started the

ACC: at the same speed, the lower the value of THW,

the lower the level of trust the driver is.

Although it is possible to quantify human-

machine relationships through trust evaluation, it is

difficult to understand the deep information, such as

the driver's distrust of ACC. What is the driver’s real

concern? What is the cause of the conflict between

the driver and ACC? Based on the above questions,

this study uses a in-depth interview method to get a

better understanding of the behaviour and thoughts of

the driver in the cut-in scenario, providing in-depth

analysis of the driver’s behaviour in this experiment.

The in-depth interview is a qualitative research

method. Researchers with specialized access skill to

Understanding Behavioural Conﬂict between the Drivers and Adaptive Cruise Control (ACC) System in Cut-in Scenario

Figure 4: Experimental environment.

to treat a certain condition to the access object, uses the

unstructured method to carry on the personal

conversation type visit, in order to reveal the hidden

secret about the specific behaviour, the motive, the

purpose, the attitude, the feeling report and discovers

its inner connection (Yuan Yue et al., 2006). The

conclusion of this study is based on this method.

The interview outline is designed to focus on the

driver's experience in cut-in scenario when driving

simulator, and interview the driver's subjective feelings

and objective behaviour during the experience. The

subjective experience includes the feeling of response

during the process of cut-in, the time and influence of

the danger consciousness, and the situational

awareness of the ACC human-machine interface.

Objective behaviour refers to the braking response of

the driver in the ACC cut-in scenario.

The selection of the participants mainly considered

the degree of understanding of the ACC system, the

experience of using ACC system, the driving

experience and several factors of age. The study

selected 8 participants, 4 males and 4 females, aged 20

to 30 years old, who had some knowledge of the ACC

system and had a 1-2year driving experience. The

selection of younger users is mainly due to the fact that

the cut-in scenario has high level requirements for the

user's situational awareness and responsiveness. The

younger users can respond faster, so that we can get

more detailed interview data.

The interview is about 30-40 minutes, a week in

length and is visited in the Automotive Media Lab of

Tongji University. During the interview, first, the

participants were introduced to the driving simulator

and the ACC. Then, the main interview purpose was

explained, and the participants were allowed to

practice driving for 10 minutes. Afterward, each

participant was shown a video of the ACC's cut-in

scenario in six different conditions, and during the

experience, the participants were able to turn the

steering wheel or brake based on their perceptions of

the driving scene, and the observers recorded their

behaviour. Finally, after the experience was over. In

the form of access, a semi-structured outline was

adopted, and the participants were able to say how they

felt in the cut-in scenario, and the facilitator pursued

the key questions based on the actions and thoughts of

the interviewees. In order to ensure the authenticity and

correctness of the interview, this study used a recording

pen to record the interview contents, and then collated

the documents and send it to the participants to

confirm.

3 RESULTS

3.1 In-depth Interview Results Analysis

The driver's prediction of the behaviour of the cut-in

vehicle’s driver in the cut-in scenario affects the take-

over behaviour. In all conditions, the driver is always

more likely to choose to brake than to feel the danger.

Of the 8 participants, 5 said that they were still taking

the brake when they were not feeling the danger

because of the predicting driver's behaviour of the side

car during driving. If the side car was too close to the

side lane, the driver might have an idea of the cutting-

in. So when the side car was near the side lane line, the

participants were already alert and ready to brake. In

one typical case, in one of the test scenarios, the car's

speed V = 50km / h, THW = 1.2s, only one of the

participants felt dangerous, but the number of people

taking the brake was 4. The reason for this is that in the

cut-in scenario, the side car continued to drive for 1-2

seconds near the side line before cutting in, which

causes the drivers to take the brake. From this we can

recognize that the driver and ACC are different in their

definition of the cut-in scenario.

Figure 5: Different definitions of driver and ACC for cut-in

scenario.

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

100

Figure 6: Human-machine interface design strategy in ACC cut-in scenario.

As shown in Fig. 5, for the driver, the cut-in scenario

is more than just the vehicle's cutting-in process, but

also the stage where the pre-sensing side car is about

to cut.

In the cut-in scenario, the driver focuses more on

the actual distance between the two cars ( the actual

distance = V THW ) than the speed or THW.

According to the researcher's observation, when the

THW is constant, the lower the speed (V=30km/h),

the greater the driver feels dangerous and brakes,

which seems to contradict the common sense that we

generally believe that the speed is higher and more

dangerous. For example, in the cut-in scenario, when

THW=1.2s, V=60km/h, the distance between the two

vehicles is 20.00m, 1 participants brake; when

V=50km/h, the distance between the two vehicles is

16.67m, 2 participants brake; and V=30km/h, the

distance between the two vehicles is 10.00m, and 5

people brake. As the two cars get closer and closer,

the more dangerous the driver is feeling. According

to the in-depth interview, two participants said that as

they got closer and closer to the front car, there was a

looming pressure, heightened awareness of the

danger, and therefore braking.

The ACC human-machine interface is less helpful

to the driver in the cut-in scenario. Of the 6

conditions, 4 of them have 2 respondents who thought

that the ACC human-machine interface was helpful,

and the remaining 2 conditions were 3 and 4

respectively. One participants said the reason why the

ACC human-machine interface wasn't helpful was: "

when I (focus) drive, and I can look out for

information. "According to the researcher's further

confirmation, the interviewee's priority in the process

was the external environment in the process of cutting

in. According to the interview, all the participants

said they were in a situation where they felt the

danger, the moment when the side car cut-in, focused

on the environment outside the car and the distance

between the car and the cut-in car. After the side car

cut-in and had been steadily moving, they would pay

attention to the ACC system and understand its status

display.

3.2 Human Machine Interface Design

Strategy

In the cut-in scenario, the performance of the ACC

human-machine interface is not satisfactory to the

driver. The main reason for the conflict is that the

driver is considering are very different from the ACC,

which is mainly reflected in three aspects: the

different definitions of the cut-in scenario, the

different between the perceived risk factors and the

stress, and the perception of the cut-in scenario.

In order to achieve a harmonious human-machine

interaction and reduce the conflict between the driver

and ACC, the ACC human-machine interface should

be designed according to the driver's cognition and

thinking model, as shown in Fig.6.

(1) Firstly, it is important to redefine the cut-in

scenario, to expand the concept of the cut-in process

to four stages: anticipative cut-in, before cut-in,

cutting in, and after cut-in. The predictive behaviour

of the side vehicle’s driver can be judged by the

distance between the side car to the side lane line.

When the side car is closer and closer to the side lane

line, the side driver is considered to have the intention

Understanding Behavioural Conﬂict between the Drivers and Adaptive Cruise Control (ACC) System in Cut-in Scenario

101

of cutting, and the driver should be informed of the

human-machine interface.

(2) Secondly, in view of the driver's perception of

danger and pressure, the ACC human-machine

interface can be combined with the multi-sensory

channels such as interface display, sound, vibration,

etc. to convey the situational pressure when the

distance between the two cars is reducing. So that

ACC human-machine interface is consistent with the

driver's psychological cognition. In the design study

of Bobbie et al., (2007), changes in size and shape

(trapezoid and triangle) was used to show the

behaviour of the ACC and the danger of the current

driving scenario, so as to consistent with the driver's

perception of the dangerous scene.

(3) Finally, in response to the driver's perception of

the ACC cut-in scenario, the ACC human-machine

interface should not only inform the driver of the

status and behaviour of the ACC, the road condition,

the driving environment, etc (Bobbie et al., 2015). but

also the driver should be warned in advance to help

the drivers deal with unexpected situations.

In the study of Bako (Bako et al., 2006), the

researchers also noted that ACC's early warning of

drivers could help drivers cope with dangerous

scenarios. The early warning should be gradual and

hierarchical. According to the importance of the

information and the degree of impact to safe driving,

it is divided into three levels: the first is the guide

information which about safe driving. This level uses

only the visual perception channel, displays on the

interface, plays the guiding role for the driver's

driving behaviour. In the cut-in scenario, the guidance

information should be displayed in the first two

stages: perception cut-in and before cut-in, giving the

driver an understanding of the current road conditions

and trends.

The second level is the warning information for

the possible threat of safe driving. In order to enhance

the driver's awareness of crisis, using both visual and

auditory channels. In the cut-in scenario, it

corresponds to the cut-in phase. When the cut-in

vehicle threatens to safe driving, a note is sounded,

and the warning interface is displayed on the dash

board, causing the driver to be alert. The third level is

to take over information which may cause a traffic

accident and need an emergency take over. In this

level, three sensing channels are used to attract the

attention of the driver to take over the vehicle: vision,

hearing and touch. In the cut-in scenario, when the

vehicle is cut-in completely and threatens to safe

driving, the user will be prompted to take over the

vehicle by displaying the danger information,

prompting the sound and shaking the steering wheel.

4 CONCLUSION

In order for the driver and ACC to achieve a

harmonious human-machine cooperation and

improve efficiency and safety of driving, we should

clearly recognize the differences and conflicts

between drivers and ACC in the cut-in scenario. The

biggest conflict is that drivers and ACC have different

definitions of the cut-in scenario, the perceptions of

danger and abilities to afford pressure of impending

danger, and the context-aware content and processes

of environmental perceptions. There are two ways to

reduce this divergence. One is to make the ACC

system's sensor performance enhanced, the control

strategy more humane, and have the ability to learn

and record the driver's driving habits and driving

behaviour, making the ACC system more and more

consistent with the driver's expectations. The second

way in which the vehicle and the ACC can detect

information and convey to the driver through the

human-machine interface by using different sensory

channels based on the degree of danger, and allow the

driver to use the strength of ACC to expand his

abilities as much as possible.

At present, the second approach, it seems, is easier

to implement and less costly. The human-machine

interface design, which is designed to reduce the

conflict between the driver and the machine, is crucial

to understanding the conflict between the two. Based

on this purpose, this research obtains first-hand

information through in-depth interviews, analyzes the

behaviour and thoughts of drivers in the cut-in

scenario, and proposes three design strategies for the

cut-in scenario, which has certain practical

significance. However, this study still has limitations.

One of the most important is the inability to fully

respond to all ages, and the behaviours and thoughts

of other drivers in the process of cut-in scenario will

enrich our understanding of this problem.

The method human-machine interface design

proposed in this study is mainly aimed at the ACC

cut-in scenario, which can be used in other driving

scenarios, and even provides some ideas for the

design of intelligent human-machine system to help

drivers better cope with various driving situations and

achieve harmonious human-machine cooperation.

These questions will be further explored in future

studies.

ACKNOWLEDGEMENTS

Thank you to the teachers and students of the Institute

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

102

of Safety Technology of Tongji University

Automobile Institute for assisting in the safety test of

ACC function.

REFERENCES

Bengler K, Zimmermann M, Bortot D, et al. Interaction

principles for cooperative human-machine systems[J].

it-Information Technology Methoden und innovative

Anwendungen der Informatik und Informationstechnik,

2012, 54(4): 157-164.

Hoc J M. Towards a cognitive approach to human–machine

cooperation in dynamic situations[J]. International

journal of human-computer studies, 2001, 54(4): 509-

540.

Annika F.L. Larsson, Katja Kircher, Jonas Andersson

Hultgren. Learning from experience: Familiarity with

ACC and responding to a cut-in situation in automated

driving. Transportation Research Part F 27 (2014) 229–

237.

Frank Flemisch, Matthias Heesen, Tobias Hesse, Johann

Kelsch, Anna Schieben, Johannes Beller. Towards a

dynamic balance between humans and automation:

authority, ability, responsibility and control in

sharedand cooperative control situations. Cogn Tech

Work (2012) 14:3–18.

Tsang-Wei Lin, Sheue-Ling Hwang, Paul A. Green. Effects

of time-gap settings of adaptive cruise control (ACC)

on driving performance and subjective acceptance in a

bus driving simulator. Safety Science 47 (2009) 620–

625.

Pengjun Zheng, Mike McDonald. Manual vs. adaptive

cruise control – Can driver’s expectation be matched?

Transportation Research Part C 13 (2005) 421–431.

Wouter J. Schakel, Cornelis M. Gorter, Joost C.F. de

Winter, Bart van Arem. Driving Characteristics and

Adaptive Cruise Control – A Naturalistic Driving

Study. IEEE intelligent transportation systems

magazine, summer 2017,17-24.

Bobbie D. Seppelt, John D. Lee. Modeling driver response

to imperfect vehicle control automation. Procedia

Manufacturing 3 (2015) 2621 – 2628.

Brian Tsang-Wei Lin, Sheue-Ling Hwang. Effect

prediction of time-gaps for adaptive cruise control

(ACC) and in-vehicle tasks on bus driver performance.

Safety Science 50 (2012) 68–75.

Pacaux-Lemoine M P, Trentesaux D, Rey G Z, et al.

Designing intelligent manufacturing systems through

Human-Machine Cooperation principles: A human-

centered approach[J]. Computers & Industrial

Engineering, 2017, 111: 581-595.

Solvang B, Sziebig G, Korondi P. Shop-floor architecture

for effective human-machine and inter-machine

interaction[J]. Acta Polytechnica Hungarica, 2012, 9(1):

183-201.

Oborski P. Man-machine interactions in advanced

manufacturing systems[J]. The International Journal of

Advanced Manufacturing Technology, 2004, 23(3-4):

227-232.

Pacaux-Lemoine M P, Trentesaux D, Rey G Z, et al.

Designing intelligent manufacturing systems through

Human-Machine Cooperation principles: A human-

centered approach[J]. Computers & Industrial

Engineering, 2017, 111: 581-595.

Yuan Yue, Fan Wen, Chen Xiao-Li. Deep communication:

How to Make Friends with Strangers. Beijing: China

Machine PRESS，2006.

Bobbie D. Seppelt, John D. Lee. Making adaptive cruise

control (ACC) limits visible. Int. J. Human-Computer

Studies 65 (2007) 192–205.

Bobbie D. Seppelt, John D. Lee. Modeling driver response

to imperfect vehicle control automation. Procedia

Manufacturing 3 (2015) 2621 – 2628.

Bako Rajaonah, Franoise Anceaux, Nicolas Tricot, Marie-

Pierre Pacaux-Lemoine. Trust, cognitive control, and

control: the case of drivers using an Auto-Adaptive

Cruise Control. Proceedings of the 13th Eurpoean

conference on Cognitive ergonomics: trust and control

in complex socio-technical systems. Zurich,

Switzerland, 2006:17-24.

Understanding Behavioural Conﬂict between the Drivers and Adaptive Cruise Control (ACC) System in Cut-in Scenario

103