Promoting Active Interaction Between Autonomous Vehicles and Their Human Drivers

An automated driving system based on the idea of “collaborative steering” has been developed by scientists from École polytechnique fédérale de Lausanne (EPFL) and JTEKT Corporation.

This system is targeted to increase efficiency, transportation safety, and ease by promoting active interaction between autonomous vehicles and their human drivers.

Already, autonomous driving technologies have been combined into several mass-produced vehicles, thereby offering human drivers steering assistance.

However, the minimal data available on automated driving safety shows that placing too much control of a vehicle in the hands of automation could result in more harm than good, as detachment by human drivers could increase the risk of accidents.

Current vehicles on the market are either manual or automated, and there is no clear way of making their control a truly shared experience. This is dangerous, because it tends to lead to driver over-reliance on automation.

Jürg Schiffmann, Head, Laboratory for Applied Mechanical Design, School of Engineering, École polytechnique fédérale de Lausanne

Now, scientists from the laboratory have partnered with Japanese steering system supplier JTEKT Corporation to engineer and successfully road-test a haptics-based automated driving system.

Various modes of human-robot interaction have been combined by this system. The scientists believe that their method will not just help in increasing the safety of automated driving, but also social acceptance of it.

“This research was based on the idea that automation systems should adapt to human drivers, and not vice versa,” states EPFL Ph.D. student and JTEKT researcher Tomohiro Nakade, who is also the first author of a recent paper explaining the system published in the Nature journal Communications Engineering.

Nakade adds that a good metaphor available for the new system could be drawn from a transportation mode that predates automation: “A vehicle must be open to negotiation with a human driver, just as a horseback rider conveys his or her intention to the horse through the reins.”

Interaction, Arbitration, and Inclusion

Contrary to present automated driving systems, which use only cameras for sensory input, the scientists’ more holistic method combines information from a car’s steering column. Also, it promotes constant engagement between drivers and automation, in contrast to present automated systems, which are normally either switched on or off.

In automation in general, when humans are just monitoring a system but not actively involved, they lose the ability to react. That’s why we wanted to actively improve driver engagement through automation.

Robert Fuchs, Ph.D. Student, R&D General Manager, JTEKT Corporation

As a result of the three functionalities, arbitration, interaction, and inclusion, it was possible for the researchers to achieve this.

Initially, the system differentiates between four various types of human-robot interaction: competition (human and automation activities are in opposition), collaboration (human and automation assist one another in achieving different goals), cooperation (the automation supports the human in achieving a goal); and coactivity (the human and automation have different goals, but their actions impact one another).

Then, as the vehicle is operated by the driver, the system arbitrates or shifts between various interaction modes based on the evolving situation on the road. For instance, the car might switch from collaboration to competition mode to prevent an abrupt collision threat.

Eventually, and still within the same control framework, the system has an “inclusion” function: it recomputes the trajectory of the vehicle whenever the driver interferes—by turning the steering wheel, for instance—instead of perceiving it as an override and switching off.

A Practical Solution

For their system to be tested, the scientists came up with experiments that involved a simulated virtual driver and a human driver utilizing a detached power steering system, a full-driving simulator, and even field tests with an altered test vehicle.

The field tests were performed with the participation of five drivers on a JTEKT test course in Japan’s Mie prefecture, by linking the scientists’ system to a standard sedan through an external controller.

The scientists particularly tested drivers’ experiences of steering smoothness and lane-changing comfort, and their outcomes verified the considerable potential of the system for increasing ease and decreasing effort for drivers via collaborative steering.

This is a very practical concept—it’s not just research for research’s sake. It’s also a beautiful example of a fruitful partnership between our lab and JTEKT, with whom EPFL has collaborated since 1998.

Jürg Schiffmann, Head, Laboratory for Applied Mechanical Design, School of Engineering, École polytechnique fédérale de Lausanne

Schiffmann adds the fact that the software-based system could be combined into standard mass-produced cars in the absence of any special equipment.

Journal Reference:

Nadake, T., et al. (2023) Haptics based multi-level collaborative steering control for automated driving. communications engineering. doi.org/10.1038/s44172-022-00051-2.

Source: https://www.epfl.ch/en/