Understanding the supervision activity to design a non-existent control system for automated driving through prospective ergonomics

Abstract

Just like the progressive evolution of industrial processes towards highly automated systems, today we see a similar evolution in the mobility sector with the development of autonomous vehicles. These evolutions have and will change the work activity of human from "realization" to "supervision". Due to the progressive increase in autonomy, the SAE (Society Of Automotive Engineers) has created a categorization in 6 levels (from 0 to 5, where 5 corresponding to a fully autonomous vehicle). As long as automated vehicles are not able to manage all driving situations (SAE levels 2, 3 and 4), the human operator remains in the control loop and cooperates with the autonomous system. In this context of progressive deployment and as it has been the case in other transportation modes such as aviation and railways, we anticipate the design of a centralized supervision center for fleet of autonomous vehicles. Although the literature highlights many bias in the relationship between humans and automation, the role of the supervision would be to secure the operation (by man's ability to recover a complex situation) of the vehicle by anticipating incidents (e.g.: support the driver-system relationship, like a traffic controller would do for pilots), while guaranteeing reliability (management of system failures) as well as the regularity of the network.To propose a specification of the high-level functions of a no-existing system, we deploy the approach of prospective ergonomics. This approach "consists in anticipating future needs, uses and behaviors or in building future needs in order to create processes, products or services that are well adapted to them" (Brangier & Robert, 2014). This recent modality of ergonomics intervention is based on the prospective foundation, [naturally] on the ergonomic foundation as well as on the creative foundation. We have set up a methodology divided into two phases.The first phase brings together the prospective and ergonomic foundations. We take up the work of Daniellou (1992) who proposes to be interested in situations which present strong similarities with the system to be designed, which he calls "reference situations". We have identified 8 sectors of activity in the spectrum of supervision such as aviation, bus, railway or nuclear. So far, we have been able to integrate four reference situations that have allowed us to identify seven components around supervision (safety, infrastructure, hardware, degree of automation, software, system organization, human factor). These elements are essential to the understanding of the supervision panorama.From the different situations observed we prepare the second phase of our methodology which corresponds to the creative foundation. Through expert staff we aim to bring elements of identification of the best method to help in the projection of the future. We will therefore compare the production of ideas between a group that will be able to base itself on the syntheses of the observed reference situations and a control group.The ambition of this work is to ameliorate the anthropocentric approach of long-term design.

Keywords: Prospective ergonomic, Autonomous véhicule, Supervision, Human Factors

DOI: 10.54941/ahfe1002777