Automated driving systems (ADSs) can be regarded as specific types of robots...

They act with different levels of autonomy based on input data from a multitude of sensors and have, as automated agents, an effect on their physical environment. Thus, we argue that ADSs are "robots on wheels". A good starting point for discussions on commonalities and distinctions between ADSs and robots are Asimov's laws [3] - three rules devised by Isaac Asimov (science fiction author) in 1942. The laws, quoted as being from the "Handbook of Robotics, 56th Edition, 2058" are:

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  2. A robot must obey the orders given it by human beings except where such orders would conflict with the first Law.
  3. A robot must protect its own existence as long as such protection does not conflict with the first or second laws.
These rules have been discussed for a long time in HRI and are valuable starting points of ADS design. Most importantly, these three rules are not isolated and need to be integrated carefully.

Differences between automated cars and traditional robots

The design of ADSs has much in common with the design of other robots typically investigated in human-robot interaction research. However, important differences exist.

One of the main differentiating characteristics between ADSs and traditional robots is, that humans who (co-)operate ADSs are sitting inside that vehicle instead of operating it from the outside.
Secondly, when an ADS is carrying a passenger, the vehicle can no longer choose to violate rule 3 in order to uphold rule 1.

After all, if the vehicle crashes in order to save the life of someone outside the vehicle, it will possibly harm its passenger, thus automatically violating rule 1. Thus, we can see that in automated driving, there is no "right" choice to make in many situations when we apply only these rules. Therefore, findings and regulations in HRI need to be extended and thoroughly discussed to accommodate the design of ADSs. In other words, automated driving requires its specific decision-making models and rules to be developed and validated. On the way to build such complete models/rules, a number of problems with unclear solutions will show up. Meanwhile, there may also be existing HRI methods that can be adapted to solve intermittent problems. We believe that the knowledge exchange between HRI in traditional robotics and automated vehicles has a high potential and both communities can learn from each other. For example, in industrial HRI context, trust is a very important component. The same is true in automated driving. Approaches regarding factors that influence trustworthiness in technology and how to design for trust have been developed in both traditional HRI and automated driving communities. We will investigate and compare approaches among different types of HRI, with the goal to share, expand, and improve the knowledge regarding trust design. Another issue in both contexts is hand-over of control and locus of control. In both communities, these issues have been discussed intensely. Solutions of how to hand-over the control from an ADS to the driver and vice versa are also interesting from the human-robot interaction perspective...