ESR8 – Human Factors in AI-based Automation Design

Popular scientific abstract

The world of transportation is undergoing a transformation thanks to advancements in automated technology, which promise to reduce road fatalities and improve mobility for both individuals and society. While some autonomous vehicle (AV) technology is already in use on our roads, there has been insufficient consideration given to the human factors involved in AV development. Understanding how humans interact with AVs, both inside and outside the vehicle, is essential for their success and widespread acceptance. Solid knowledge about human factors is crucial throughout the entire AV development process to ensure the safety and trust of users.

However, the recent trend in the automotive industry towards agile development methodologies makes it difficult to incorporate human factors knowledge into requirements. This study aims to investigate how to effectively bring human factors requirements to AV developers in large-scale agile development and provide guidelines and methodologies for doing so.

Who am I?

I am Amna Pir Muhammad, a Ph.D. student at the Chalmers | University of Gothenburg. Currently, I am working under the supervision of Dr. Eric Knauss and Dr. Jonas Bargman on a Marie Skłodowska-Curie Action Innovative Training Network project.

My research focuses on bringing Human Factors (HF) knowledge to AV developers. I believe it is vital to incorporate HF knowledge into AV development to make it more reliable and efficient. Before my Ph.D., I specialized in Software Engineering and worked as a lecturer at the Comsats University of Islamabad.

My passion lies in shaping the relationship between humans and AI-based autonomous systems, particularly in the fields of autonomous vehicle development, large-scale agile development, requirements engineering, and human factors.

Supervisors

Eric Knauss:                         eric.knauss@cse.gu.se

Jonas Bärgman:                  jonas.bargman@chalmers.se

Alessia Knauss:                  alessia.knauss@zenseact.com

Introduction and Motivation

By moving towards eliminating human-driver-caused crashes, automated Vehicles (AVs) promise a number of benefits: fewer accidents, injuries, and deaths, as well as enabling drivers to engage in other activities while in the car [1, 2, 3]. Due to this promise, the automotive industry is currently competing to develop and market AVs with increasing levels of automation [4], ranging from specific automated functions in Advanced Driver Assistance Systems (ADAS) that can support the driver in the driving task to fully autonomous vehicles that take over all driving tasks—at least under specific conditions (Operational Design Domains; ODDs; [5])—that do not require supervision. AVs are always software-intense and often rely on artificial intelligence (AI). Consequently, AVs are complex systems that require careful consideration in their design.

On the other hand, in spite of all their benefits, AVs also pose various challenges to humans, such as over-trust and over-reliance, extra workload on humans, or driver engagement and re-engagement [6, 7]. The challenges are not limited to drivers; other road users who interact with AVs will also be affected. Actually, unsafe and non-human-centered interactions between AVs and drivers and other road users can substantially reduce the benefits of AVs, affecting humans both inside and outside the vehicle.

To overcome the issues of managing human factors in vehicle automation (see, e.g., [8] for examples of automation failures) and achieve the full potential of automation, human factors researchers strongly advocate considering knowledge about human factors when designing AVs [9, 10, 11]. This field, the study of human factors, involves researching human capabilities and limitations and other human characteristics and applying the findings to the design of systems to improve performance, safety, and comfort [12].

In order to include human factors when designing automation, researchers recommend incorporating human factors knowledge into the early stages of development [13, 14, 15]. Traditionally such knowledge has been included in system requirements that have been specified up-front and form the foundation of subsequent design work [16]. The process of eliciting, analyzing, documenting, and validating the requirements during the engineering process is called Requirements Engineering (RE) [17]. However, the recent trend in the automotive industry to adopt agile development methodologies changes the role of RE significantly. Agile development methodologies are a collection of approaches based on incremental and iterative development in which self-organizing and cross-functional teams work together to generate requirements and functions [18]. Agile methods aim to deliver faster in less time to market since, due to competition, developers want to deliver faster; they focus on technical details and often neglect others, such as those provided by human factors. Moreover, because agile methodologies do not focus on the processes, RE processes are not well integrated with agile methodologies and face different challenges [19].

Without a clear role for RE in agile development, there is a risk that introducing human factors knowledge as requirements will be difficult. This risk is increased by the lack of empirical research on how to include human factors knowledge in agile development; practitioners struggle with a lack of clear guidelines.

Introduction video from 2020

Objectives

The main objectives of this research are

  1. To understand and describe how AI-based AV design can account for human factors (for example, how human comfort-zone boundaries and acceptance can be considered when designing AI-based strategies for AV control)
  2. To develop methods to incorporate results from studies on acceptance which will improve AI-based AV-designs
  3. To identify and integrate disparate requirements from AV human factors researchers and designers of AI-based AV control in order to improve road-user acceptance, AV transparency, and vehicle safety

Results

The expected results are:

  1. Methods and guidelines for incorporating explicit human factors knowledge into AI-based AV designs (requirements)
  2. Recommendations on how information should be communicated between AV human factors researchers and AI-based AV designers to improve road-user acceptance, AV transparency, and vehicle safety
  3. Recommendations on how to optimize communication (particularly knowledge transfer) between human factors researcher and AI-based AV designers

My Publications

  • Muhammad, A. P., Knauss, E., Batsaikhan, O., Haskouri, N. E., Lin, Y. C., & Knauss, A. (2022, November). Defining Requirements Strategies in Agile: A Design Science Research Study. In Product-Focused Software Process Improvement: 23rd International Conference, PROFES 2022, Jyväskylä, Finland, November 21–23, 2022, Proceedings (pp. 73-89). Cham: Springer International Publishing.
  • Muhammad, A. P. (2022). Managing Human Factors and Requirements in Agile Development of Automated Vehicles: An Exploration (Licentiate dissertation, Chalmers Tekniska Hogskola (Sweden)).
  • Heyn, H. M., Knauss, E., Muhammad, A. P., Eriksson, O., Linder, J., Subbiah, P., … & Tungal, S. (2021, May). Requirement engineering challenges for ai-intense systems development. In 2021 IEEE/ACM 1st Workshop on AI Engineering-Software Engineering for AI (WAIN) (pp. 89-96). IEEE.
  • Muhammad, A. P. (2021). Methods and Guidelines for Incorporating Human Factors Requirements in Automated Vehicles Development. In REFSQ Workshops.
  • Sultan, Muniba, Amna Pir, and Nazir Ahmad Zafar. “UML-based formal model of a smart transformer power system.” International Journal of Advanced Computer Science and Applications 8, no. 11 (2017).
  • Hassan, T., Hassan, S., Muhammad, A. P., & Yar, M. A. (2016). Semantic Technology and Ontology. International Journal of Computer Science and Information Security, 14(5), 212.
  • Pir, Amna, M. Usman Akram, and Muazzam A. Khan. “Internet of Things based context awareness architectural framework for HMIS.” In 2015 17th International Conference on E-Health Networking, Application & Services (HealthCom), pp. 55-60. IEEE, 2015.
  • Muhammad, A. P., Akram, M. U., & Khan, M. A. (2015, December). Survey-based analysis of Internet of Things based architectural framework for a hospital management system. In 2015 13th International Conference on Frontiers of Information Technology (FIT) (pp. 271-276). IEEE

References

  1. D. J. Fagnant and K. Kockelman, “Preparing a nation for autonomous vehicles: opportunities, barriers and policy recommendations,” Transportation Research Part A: Policy and Practice, vol. 77, pp. 167–181, 2015.
  2. W. D. Montgomery, “Public and private benefits of autonomous vehicles,” 2018.
  3. L3Pilot, “L3pilot driving automation,” 2022, accessed on 15 Nov, 2022. [Online]. Available: https://l3pilot.eu/downloads
  4. SAE, “SAE J3016:201806 – SURFACE VEHICLE RECOMMENDED PRACTICE – Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles,” 2018.
  5. N. H. T. S. Administration et al., “Automated driving systems 2.0: A vision for safety,” Washington, DC: US Department of Transportation, DOT HS, vol. 812, p. 442, 2017.
  6. C. E. Billings, Aviation automation: The search for a human-centered approach. CRC Press, 2018.
  7. N. Merat, A. H. Jamson, F. C. Lai, M. Daly, and O. M. Carsten, “Transition to manual: Driver behaviour when resuming control from a highly automated vehicle,” Transportation research part F: traffic psychology and behaviour, vol. 27, pp. 274–282, 2014.
  8. W. Biever, L. Angell, and S. Seaman, “Automated driving system collisions: early lessons,” Human factors, vol. 62, no. 2, pp. 249–259, 2020.
  9. P. A. Hancock, “Some pitfalls in the promises of automated and autonomous vehicles,” Ergonomics :1, 2019.
  10. C. Wickens, J. Lee, Y. Liu, and S. Gordon-Becker, Designing for People: An Introduction to Human Factors Engineering. CreateSpace, Charleston, 2018.
  11. J. Navarro, “A state of science on highly automated driving,” Theoretical Issues in Ergonomics Science, vol. 20, no. 3, pp. 366–396, 2019.
  12. H. Factors and E. Society, “Definitions of human factors and ergonomics,” 2021, accessed on 17 Feb, 2021. [Online]. Available: https://www.hfes. org/About-HFES/What-is-Human-Factors-and-Ergonomics
  13. M. H. Calp and M. A. Akcayol, “The importance of human computer interaction in the development process of software projects,” arXiv preprint arXiv:1902.02757, 2019.
  14. Z. K. Chua and K. M. Feigh, “Integrating human factors principles into systems engineering,” in 2011 IEEE/AIAA 30th Digital Avionics Systems Conference. IEEE, 2011, pp. 6A1–1.
  15. E. H˚akansson and E. Bjarnason, “Including human factors and ergonomics in requirements engineering for digital work environments,” in 2020 IEEE First International Workshop on Requirements Engineering for Well-Being, Aging, and Health (REWBAH). IEEE, 2020, pp. 57–66.
  16. W. W. Royce, “Managing the development of large software systems: concepts and techniques,” in Proceedings of the 9th international conference on Software Engineering, 1987, pp. 328–338.
  17. G. Kotonya and I. Sommerville, Requirements engineering: processes and techniques. Wiley Publishing, 1998.
  18. A. Moniruzzaman and D. S. A. Hossain, “Comparative study on agile software development methodologies,” arXiv preprint arXiv:1307.3356, 2013.
  19. R. Kasauli, E. Knauss, J. Horkoff, G. Liebel, and F. G. de Oliveira Neto, “Requirements engineering challenges and practices in large-scale agile system development,” Journal of Systems and Software, vol. 172, p. 110851, 2021.

Contact information

Amna Pir
Department of Computer Science and Engineering 
Chalmers | University of Gothenburg
email: amnap@chalmers.se
LinkedIn, Google Scholar , Twitter