Cooperation Relationship of Human-Machine Teaming on

Co-Driving for Automous Vehicle

Zhou Shijie

1,2

, Liu Hailing

, Su Yuhang

, Zhang Wendi

and Ma Jun

3,*

College of Applied Science and Technology, Hainan University, Hainan, China

Department of User Experience, Beijing Electric Vehicle CO.LTD, Beijing, China

Tongji University, Shanghai, China

Keywords: Human-Machine Teaming, Autonomous Vehicle, Co-Driving.

Abstract: In human-machine symbiotic intelligent system, the relationship between man and machine is a kind of team

cooperation. With the maturity of autonomous driving technology, intelligent vehicles have developed from

a driving assistance tool into an “autonomous intelligent agent” with certain cognition, independent execution

and self-adaptation. We will no longer see the automatic vehicle as a "machine", but as a "partner". Therefore,

based on the cooperation mechanism of "Human-Human Teaming (HHT)", we propose the psychological

cognitive framework of Human-Machine Teaming (HMT). They work together to complete co-driving tasks

in the process of bi-directional trust, situational awareness and share control right. This reflects the deep

integration of biological intelligence represented by human brain (cognitive information processing ability)

and machine intelligence represented by computer technology (industrial artificial intelligence) to achieve the

intelligent complementarity. Human operator and intelligent system cooperate with each other at multiple

intelligence levels, such as perception, analysis, reasoning and decision-making, so as to realize the overall

match and the effective cooperation in human-machine group.

1 INTRODUCTION

With the advances machine learning and artificial

intelligence, automation systems can independently

perform some scene tasks without human

intervention (Kaber, 2018)，which means it having

certain adaptive abilities and owning a greater degree

of autonomy (The Atlantic, 2013). Machine based on

intelligent system is developing from an auxiliary tool

supporting human operation to an autonomous

intelligent agent with certain cognitive, independent

execution and self-adaptation abilities, which has

behaviors similar to human beings to a certain extent.

Early Automation typically employ logic-based

programming to accomplish tasks with little or no

human intervention, and it is widely defined as

"functional machine execution. More specifically, it

is "a technique for actively selecting data,

transforming information, making decisions, or

controlling processes" (Lee and See, 2004). While the

https://orcid.org/0000-0003-2975-9486

https://orcid.org/0000-0002-1629-1530

Corresponding author

autonomous system is based on the computational

intelligence and learning algorithm, which evolves

according to the input of operation and better adapts

to the constantly changing situation in order to

achieve the goals without manual intervention

(Endsley, 2017). However, achieving totally

autonomy is quite difficult. Therefore, for the most

autonomous systems will exist for a long time with

some degree of semi-autonomy, that is, certain

aspects of the system develop to into autonomy, but

human must be in circle and have the ultimate

decision power.

Recent examples of such highly autonomous

technology is that self-driving vehicles are already

beginning to propagate through our society. With the

maturity of automatic driving technique, the

intelligence and autonomy of vehicle system is also

improving. Different from the automated system

which only serves as a driving assistance tool in the

past, the autonomous system can become a

156

Shijie, Z., Hailing, L., Yuhang, S., Wendi, Z. and Jun, M.

Cooperation Relationship of Human-Machine Teaming on Co-Driving for Automous Vehicle.

DOI: 10.5220/0011916100003612

In Proceedings of the 3rd International Symposium on Automation, Information and Computing (ISAIC 2022), pages 156-161

ISBN: 978-989-758-622-4; ISSN: 2975-9463

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

“collaborator” with human cooperation to share tasks

and control rights. Modern interactions with

technology are increasingly moving away from

simple human use of machines as tools to the

establishment of human relationships with

autonomous entities (Kaber, 2018). As a result, we

will no longer see the automatic vehicle as a "tool,"

but as a “teammate”to share and complete team tasks

together.

Ideally, the combination of human and autonomy

system should result in an efficient team that

successfully promote the team performance and

avoids mistakes usually made by a single decision

maker alone (Mosier and Skitka, 1996). Therefore, it

is necessary to explore the HMT cooperation in order

to create a bidirectional fusion system that can truly

improve the quality of human-machine joint

performance. Thus, the design of such an autonomous

support system must base on a thorough analysis of

the psychological framework of the human-machine

team in the perception 、 cognition and decision-

making process.

2 PSYCHOLOGICAL

FRAMEWORK OF HMT

Given that we have identified some of the differences

between automation and autonomous systems, a

fundamentally different perspective on the human-

machine relationship is necessary. Intelligent

technology and its autonomy characteristics promote

the transformation of the human- machine

collaboration system, thus it can be seen that a new

type of human-machine cooperation has emerged,

which has evolved into a team-mate relationship—

Human-Machine Teaming (HMT) cooperation, or it

is called Human-Autonomy Teaming cooperation

(Kim and Hinds, 2006; Xu, 2021). We define human-

machine teaming (HMT) as “the dynamic

arrangement of humans and cyber-physical elements

into a team structure that capitalizes on the respective

strengths of each while circumventing their

respective limitations in pursuit of shared goals.

Some researches into the essence of relationship

between human and autonomy shows that users tend

to apply human-human interaction norms to their

interactions with "intelligent machines". At this point,

the role of the human operator has changed from a

primary controller to an active teammate who

complete together on the tasks. The machine

changing from automated (requiring human

supervision) to autonomous (not requiring human

supervision), thus we can get a recognition that

machines have greater autonomous on the decision-

making and control right. In the foreseeable future, it

is imperative that advances should be made in

effective human teaming with autonomous systems.

The paradigm of human-autonomy interaction

should adopt human-human model as its initial

standard, and take into account the autonomous

characteristics of intelligent machines, so as to further

extend the application to the human-autonomy

relationship (Shively et al., 2017). Based on the

certain key features of the mature Human-Human

Teaming theory, combining with the functions

(cognitive, control and perceptual aids ) of the

autonomous systems (Schaefer et al., 2014)，some

basic principles are established for Human- Machine

Teaming ergonomics research: bi-directional

communication and trust, shared intention and

situational awareness, controllable workload and

decision making.

The current psychological research on HMT is

mainly carried out in some industrial fields. For

examples, administers − air autonomy system

teaming in air traffic management system (Ho et al.,

2017), operator − intelligent robot teaming in special

environments (Kistan et al., 2018), pilot – aircraft

autonomous system teaming (Calhoun et al., 2018),

drivers − advanced autonomous vehicles systems

teaming (Brandt et al., 2017). The researches focuses

on some basic issues in engineering psychology,

including theoretical framework, the characteristics

of human-machine cooperation and the specific

cooperation content: bi-directional trust, situational

awareness and shared decision making. Therefore,

the following content will adopt the psychological

framework of HHT theory to analyze and summarize

the content of HMT cooperation.

Replacing human drivers with intelligent vehicles

is the goal of the development of autonomous driving

system. SAE(2019) divides the advanced driver

assistance systems into five "automation" levels (L1-

L5). However, due to the current technology, traffic

environment, policies and regulations, public

acceptance, moral principles and other reasons, there

is still a long way to go for fully autonomous driving.

In the foreseeable future for quite a long time, human

drivers and autonomous system need to control the

intelligent vehicles and complete the driving task

together, while the separation of both sides can

independently, which means the relationship between

the driver and the autonomous vehicles is more of a

team cooperation (Changfu et al., 2021).

Cooperation Relationship of Human-Machine Teaming on Co-Driving for Automous Vehicle

157

3 HUMAN-MACHINE CO-

DRIVING FOR ADAS

The purpose of this study is to explore the new

paradigm of human-machine cooperation, this paper

explores the specific content of co-driving from the

perspectives of human-machine mutual trust, shared

situational awareness and control share control right

of autonomous vehicles. The main contributions of

this research are as follows.

3.1 Bi-Directional Trust

As complex automation is now being produced and

continues to approach increasingly more advanced

intelligent, autonomous systems provide higher-level

functionality as mature team members. The value of

any such system resides not In the total replacement

of a human controller but rather in the capacity for

human -- machine collaboration. This requires the

establishment of effective team relationships, in

which trust is a crucial dimension. It is not only the

basic feature of Human-Human Teaming, but also a

key factor in regulating the relationship between

human and machine (Kistan et al., 2018).

In 1994, Muir extended Barber's definition of

interpersonal trust to human-machine relationship,

clarifying the connotation and dynamic nature of

automated trust in a complex and hierarchical

supervised control environment (Calhoun et al.,

2018).For a human-machine team to accomplish its

goals, the human operator must trust the machine

partner would protect the interests and welfare of the

whole team. This is the concept of trust in automation,

which is a primary issue affecting the effectiveness of

human-machine systems, especially when it relates to

safety, performance and utilization (Changfu et al.,

2021).Empirical studies have confirmed its critical

significance such as in high-risk situations (Brandt et

al., 2017).

In HMT, the trust between two cognitive agents

(human operator and autonomy) is bidirectional, and

whether the both sides maintain an appropriate level

of trust for each other will directly affect the team

performance (Navarro, 2019). For example,

numerous human factors studies have clearly shown

that human operator over-trust or lack of trust in

automation would have the catastrophic

consequences during real-world incidents

(Parasuraman and Riley, 1997) . On the contrary, for

the intelligent vehicles, if the driver in autonomous

driving mode leaves the steering wheel with both

hands, the agent can judge that the current status of

human operator is untrustworthy, so as to activate

some alarm method to ensure the human-in- the- loop.

In the study of man-machine teaming, how to

improve the bi-directional communication and trust is

the key to develop team trust. Mercado found that

when interacting with an intelligent planning agent,

operator performance and trust in the agent increased

as a function of agent transparency level (Mercado et

al., 2016). Moreover Chen and her colleagues have

developed the Agent Transparency model, an

effective tool to promote and calibrate team trust, to

facilitate human operators’ understanding of the

agent’s intent, logic and expected outcomes in order

to modulate their reliance on the agent (Chen et al.,

2018; Chen et al., 2014).

3.2 Situational Awareness

In the real-time changing traffic environment, many

driving decisions are required across a fairly narrow

space of time, and tasks are dependent on an ongoing

up-to-date analysis of the environment. Therefore

obtaining and maintaining good situational awareness

is a foundation for ensuring that the operator has

adequate knowledge and understanding of the

surrounding environment, which is central to

effective decision making and control in dynamic

systems. It is formally defined as "the perception of

the elements in the environment within a volume of

time And space, the Comprehension of their meaning

and the projection of their status in the near future "

(Endsley, 1998). For a multi-member team,

situational data perception comes from the collection

of every cognitive subjects, that is, team SA can be

thought of as the overlap between the aware-ness held

by individuals through communication among team

members.

Thus, the quality SA of team members' (as a state

of knowledge) can be used as an indicator of team

coordination or system effectiveness. Various

theories have been proposed for the way in which a

team creates SA, which mainly includes shared

situational awareness and distributed situational

awareness (Stanton, 2016). Shared SA, which is

assumed that an "internal" representation of each

individual's key information is shared with all other

members of the team, so that all of them hold the same

knowledge. While distributed SA considers

situational awareness is the accumulation of

information related to a particular function. In other

words, ‘no one member has the overall SA, rather it

is distributed around the system (Salmon et al., 2006).

The preliminary study of Kitchin and Baber shows

that the performance of distributed SA in team

ISAIC 2022 - International Symposium on Automation, Information and Computing

158

cooperation is higher than that of shared SA (Kitchin

and Baber, 2016), because the former can play to the

characteristics of each cognitive agent ， which

paying attention to the respective advantage on

operation and information of situation required by

both two agents.

In actual operation ， "human perception" uses

multi-modal measurements to assess the physical

function characteristics of drivers (including

distraction, fatigue, cognitive emotional state, etc.).

"Agent perception" uses sensors and computational

models to assess the human behavior, system and

environment SA (including sensing systems such as

cameras and lidar equipment, intelligent interactive

interfaces such as voice input ， vision display,

etc.).Technological artefacts, as well as human

operators, could actually represent different aspects

of SA, of which can present integration challenges.

Therefore, some studies have proposed to adopt the

theoretical framework of cooperative cognitive

system (Hollnagel and Woods, 2005). Human

operator and autonomous cognitive agent in HMT can

be regarded as two cooperative cognitive agents in the

same cognitive cooperative system. Overall team SA

can be conceived as the degree to which every team

member possesses the SA required for his or her

responsibilities. Human biological intelligence and

machine intelligence can realize intelligent

complementarity through deep integration, so as to

achieve goals that cannot be achieved by each

individually and support human-machine cooperation

effectively.

3.3 Shared Control Right

In the traditional human-computer interaction, the

machine based on computer technology serves people

as a decision support tool, while the human-machine

teaming emphasizes the sharing of decision-making

between human operator and intelligent agents.

Effective HMT should allow the sharing control right

between both two intelligent agents at various kinds

of task, function and system, provide appropriate

degree of autonomy level to choose the mode of take-

over or delivery, so as to achieve the best match and

cooperation between human and intelligent machine

in system design.

For example, in the field of autonomous vehicles,

Muslim and Itoh (2019) proposed a human-centered

approach of autonomous vehicles by an adaptive

control strategy for switching control permissions

between human and machine. For the final decision

control of the intelligent system, the human-centered

AI design concept model emphasizes that AI is to

enhance the ability of human rather than replace

human, so the guarantor should always be in the

circle, and the guarantor should be the final decision

controller of this intelligent system. Through the deep

learning of the self-optimization decision algorithm

of the system, the coordination between human and

autonomous driving system is maximized，so that

the driver can accurately optimize the driving

decision in the dynamic changing traffic

environment.

In terms of practical application in the field of

automation, studies have shown that dynamic

autonomy is a strategic solution to improve the

efficiency of human-machine cooperation in the field

of human-machine interaction. That is to say, the

intelligent system would evaluate whether human

operators can accomplish the task objectives, and

then decide whether to wait for the operation

instructions of human beings or respond

autonomously by intelligent agent (

Hollnagel and

Woods, 2005

; Hardian et al., 2006). According to the

driver's state and traffic situation, it dynamically

adjusts the allocation of team tasks and control

authority in HMT, in which human is engaged in

strategic, planning and decision- making, while the

intelligent agent is responsible for the specific

operational tasks. This group decision-making

process is the contribution of the decision support as

an auxiliary to the team collaboration. Take high-

grade autonomous driving vehicles for example,

when the driver is in a low load state, the system

encourages the human to operate manually to keep

the driver effectively monitoring and be in the circle.

When the driver is in a high load operation, the

system assists human to control the car, and the

human-machine interface should highlight the

display of the current road environment and driving

target, so that the autonomous driving vehicles can

effectively perform the driving task.

4 CONCLUSIONS

Nowadays autonomous characteristics of intelligent

technology bring a new kind of human-machine

cooperative relationship—Human-Machine

Teaming. The intelligent agent is not only a tool to

support human work, but also can become a teammate

to cooperate with human. Facing of industrial

artificial intelligence problems, in addition to

resolving the conflict of human-machine for the

purpose，this paper innovatively puts forward the

psychological framework of HMT cooperation,

which uses Human-Machine Teaming cooperation

Cooperation Relationship of Human-Machine Teaming on Co-Driving for Automous Vehicle

159

mechanism to complete co-driving task in automatic

driving. This partnership is a exactly good two-way

relationship between human and machine cognitive

agents, which is active, sharing, complementary,

replaceable, adaptive, goal- driven and predictable.

The ultimate goal for researchers should be to

emulate the best functioning human–human teams in

order to achieve the best match between human and

intelligent system and effective cooperation

teamwork in group.

The research on co-driving under autonomous

driving is currently in its infancy. This paper proposes

some preliminary suggestions on bi-directional trust,

situational awareness and share control right to ensure

the overall reliability and safety of the driving process

and maximize the effectiveness of the intelligent

autonomous driving system, which also providing

inspiration for the future driving decision mechanism

and adaptive driving right transformation strategy

under the human-vehicle co-driving. It is an

innovative application of industrial psychology in the

field of human-machine symbiotic intelligence to

improve the perception, cognition, adaptability and

autonomy of the whole system, as well as guiding

theoretical and practical significance for enriching the

interaction dimension of human-machine teaming in

the future research.

REFERENCES

Kaber D B. A conceptual framework of autonomous and

automated agents [J].Theoretical Issues in Ergonomics

Science.2018, 19(4):406-430.

The Atlantic, https://www.theatlantic.com/magazine/

archive/2013/03/the-robot-will-see-you- now/309216/

Lee, J. D., & See, K. A. (2004). Trust in Automation:

Designing for Appropriate Reliance. Human Factors,

46(1), 50–80.

Endsley, M. R. (2017). From here to autonomy: lessons

learned from human–automation research. Human

factors, 59(1), 5-27.

Kaber D B.A conceptual framework of autonomous and

automated agents [J]. Theoretical Issues in Ergonomics

Science 2018, 19(4):406-430.

Mosier, K.L. and Skitka, L.J., 1996, Human decision

makers and automated decision aids: made for each

other? In Automation and Human performance: Theory

and Applications, R. Parasuraman and M. Mouloua

(Eds), pp. 201–220 (Mahwah, NJ: Lawrence Erlbaum

Associates).

Kim, T., & Hinds, P. (2006). Who should I blame? Effects

of autonomy and transparency on attributions in

human-robot interaction. In ROMAN 2006-The 15th

IEEE International Symposium on Robot and Human

Interactive Communication (pp. 80-85). IEEE.

Xu W (2021). From automation to autonomy and

autonomous vehicles: Challenges and opportunities for

human-computer interaction [J]. Interactions 26(4):49-

53.

Shively, R. J., Lachter, J., Brandt, S. L., Matessa, M.,

Battiste, V., & Johnson, W. W. (2017). Why human-

autonomy teaming?. In International conference on

applied human factors and ergonomics (pp. 3-11).

Springer, Cham.

Schaefer, K. E., Billings, D. R., Szalma, J. L., Adams, J. K.,

Sand- ers, T. L., Chen, J. Y. C., & Hancock, P. A.

(2014). A meta- analysis of factors influencing the

development of trust in automation: Implications for

human-robot interaction (Report No. ARL-TR-6984).

Aberdeen Proving Ground, MD: U.S. Army Research

Laboratory.

Ho, N., Johnson, W., Panesar, K., Wakeland, K., Sadler, G.,

Wilson, N., ... & Brandt, S. (2017, September).

Application of human-autonomy teaming to an

advanced ground station for reduced crew operations.

In 2017 IEEE/AIAA 36th Digital Avionics Systems

Conference (DASC; pp. 1-4). IEEE.

Kistan, T., Gardi, A., & Sabatini, R. (2018). Machine

learning and cognitive ergonomics in air traffic

management: Recent developments and considerations

for certification. Aerospace, 5(4), 103.

Calhoun，G. L. ， Ruff， H. A.， Behymer，K. J. ， &

Frost ， E. M. (2018) . Human-autonomy teaming

interface design considerations for multi-un- manned

vehicle control. Theoretical Issues in Ergo- nomics

Science，19(3) , 321 - 352.

Brandt, S. L., Lachter, J., Russell, R., & Shively, R. J.

(2017). A human-autonomy teaming approach for a

flight-following task. In International Conference on

Applied Human Factors and Ergonomics (pp. 12-22).

Springer, Cham.

Zong Changfu, Dai Changhua, & Zhang Dong. (2021).

Research status and development trend of human-

machine co-driving technology for intelligent vehicles.

China Journal of Highway and Transport, 34(6), 214.

Navarro, J. (2019). A state of science on highly automated

driving. Theoretical Issues in Ergonomics Science,

20(3), 366-396.

Parasuraman, R. and Riley, V., “Humans and automation:

Use, misuse, disuse, abuse,” Human Factors 39(2), 230-

253 (1997).

Mercado, J. E., M. A. Rupp, J. Y. C. Chen, M. J. Barnes, D.

Barber, and K. Procci. 2016. “Intelligent Agent

Transparency in Human-Agent Teaming for Multi-

UxV Management.” Human Factors: The Journal of the

Human Factors and Ergonomics Society 58.3(2016):

401–415. doi: 10.1177/ 0018720815621206.

Chen, J., Lakhmani, S., Stowers, K., Selkowitz, A., Wright,

J. and Barnes, M. (2018) “Situation awareness-based

agent transparency and human-autonomy teaming

effectiveness,” Theoretical Issues in Ergonomics

Science (in press).

Chen, J., Procci, K., Boyce, M., Wright, J., Garcia, A. and

Barnes, M., "Situation awareness- based agent

ISAIC 2022 - International Symposium on Automation, Information and Computing

160

transparency," (ARL-TR-6905). U.S. Army Research

Laboratory, Aberdeen Proving Ground, MD. (2014).

Endsley, M. R. (1988, October). Design and evaluation for

situation awareness enhancement. In Proceedings of the

Human Factors Society annual meeting (Vol. 32, No. 2,

pp. 97-101). Sage CA: Los Angeles, CA: Sage

Publications.

Stanton, N. A. (2016). Distributed situation awareness.

Theoretical Issues in Ergonomics Science, 17(1), 1-7.

Salmon, P., Stanton, N., Walker, G., & Green, D. (2006).

Situation awareness measurement: A review of

applicability for C4i environments. Applied

ergonomics, 37(2), 225-238.

Kitchin, J., & Baber, C. (2016). A comparison of shared and

distributed situation awareness in teams through the use

of agent-based modelling. Theoretical Issues in

Ergonomics Science, 17(1), 8- 41.

Hollnagel, E., & Woods, D. (2005). Joint cognitive

systems: Foundations of cognitive systems engineering.

Boca Raton, FL: CRC Press. Keynote address, HCI

International, Beijing, June 2007. [26]Schermerhorn,

P., & Scheutz, M. (2009). Dynamic robot autonomy:

Investigating the effects of robot decision-making in a

human-robot team task. In Proceedings of the 2009

international conference on multi-modal interfaces (pp.

63-70).

Hardian, B., Indulska, J., & Henricksen, K. (2006, March).

Balancing autonomy and user control in contextaware

systems a survey. In Fourth Annual IEEE International

Conference on Pervasive Computing and

Communications Workshops (PERCOMW'06; pp. 6-

pp). IEEE.

Cooperation Relationship of Human-Machine Teaming on Co-Driving for Automous Vehicle

161