Meeting week 4 notes (25/9)

Present during the meeting: Luca and Marthe, Jelco (online)

Worked on during last week:

• Creating an accident situation with a car upside down to indicate an accident happened (not yet done)
• Planning a route for the AV
• Adding more agents to the environment to interact with each other. This worked, but the added pedestrian wasn't able to walk.
• Figuring out the handing over of control from AV to manual. This worked somewhat but the way to do it was very complicated and might not work during the experience. We might solve this by spawning a new car instead of changing the control. Luca will look into this for next week.
• We also need to think about our scenario more related to the theory and in relation to the learning objectives.

Revising the learning objectives.

Learning objectives:

1. Apply human-centered design methods and evaluation metrics within an automotive context.

• This is for example the colours example and associations with colours; red = bad/stop, green=good/go, blue=neutral. Also, driving behavior of the av related to other road users and trust and ethical aspects.
• We can add to the experience that the participant fills out a survey or anything about trust in what the av did/decisions the av made. The av could ask the passenger about feedback about how it went and adapt accordingly in the driving behavior. The car will tell the passenger that it will update the 'driving behavior' according to the experience. The way in which the car asks about the feedback is important, because that way, passengers can be convinced better to share their data with the aim to improve the driving behavior/efficiency.
• Other example: have the passenger choose between possible routes that the car will propose, including priorities of the passenger (time, avoid certain roads/parts, something else?) also in terms of making a trip as comfortable as possible for the passenger by playing personal favourite music, interior design, etc.
• Conclusion: the interaction between the car and the passenger is most important and should be optimized.

2. Analyze a driving task in terms of perceptual, attentional, environmental, and societal processes. --> what do humans need to do when driving and what does society think about it?

• perception: what is your position as a traffic participant and how you engage with other road users, so split between av and passenger part. perceptual is being aware of your world around you, what sensory information is coming in from around you; especially visual (road signs, traffic lights, lane markings, other road users, etc.) and auditory (horns, sirens, etc.) information in this specific context.

• attention: is where your are focussing on and what you are doing. how you focus/filter the incoming sensory information (selective vs divided attention).

• environmental: external factors like other road users communicating with you, weather conditions, road properties, traffic information (density), etc.

• societal: about acceptance and trust of AVs; how does our scenario play into this. Ethical issues as well. Expectations/regulations/norms influencing driving behavior and interaction. The example mentioned below belongs here, where signalling something could mean different things in different cultures, so this should be taken into account.

• everything that is intuitive to do for a human is extremely hard for AVs because you cannot program intuition. cultural aspects as well, like flashing your lights, how does a car interpret this, that is very hard. for example in case of when you are signalling for right of way, it could also mean you are angry at another road user. we can think about a way to implement this.

3. Analyze connections between key theories and principles within the automotive context

• go through the chapters/topics in the book and write down/discuss what we implemented from those topics/chapters into our scenario.

For every aspect, come up with an example and incorporate the practical aspect into unity.

To do's:

• Marthe: For the chapters in the book, go through every lecture and relate it to our scenario.
• Luca: look into hand over in terms of changing an AV car into a driver car, try looking into how waypoints trigger certain actions (need to think about how the car will stop), possible merge it with the countdown (for handing over control).
• Jelco: looks further into how to hand over control from AV to manual/driver (changing script), implementing game controller for vibrations, Look into the scenario building in terms of waypoints and setting up the route for the AV (make the car move).