Summary & Pictures

Emotion Garage was the most well attended workshop at Automotive UI. The workshop was planned as a design thinking workshop to explore and design to user experiences of emotions in the vehicle. The workshop consisted of several exercises exploring the ways in which emotions could be detected, the relative difficulty of detecting different emotions in different channels, identifying the most significant emotions that occur during driving, speculating on how autonomous driving will change the emotions associated with driving and a design prototyping session in which people prototyped solutions to emotionally difficult situations on the road. At the end of the workshop we recorded one minute videos reflecting on our experiences.

Session 1: Icebreaker/Sensing Channels Exercise

The first session was designed to be an icebreaker and an exploration of the relative difficulty of detecting different types of emotions across different sensing channels. Each participant was given a card that specified an emotion e.g. “Joy, Sadness, Insecurity, Embarrassment, Anger” a channel of expression, e.g. “Acoustic, Facial Expression, Gesture, and Drawing” A card looks like:

Emotion: ANGER

Channel: DRAW

Each workshop participant was given a card and was instructed to keep it secret. They went up to another participant and expressed their emotion through the channel specified. If the other participant was able to guess the emotion the guesser got a point. You could only guess once. Emotions like Anger and Joy were easy to guess. Insecurity was more difficult, especially in when expressed in modes like “drawing.” The exercise was designed to have people mix and mingle but also to start them thinking about the complexity of emotions and how different emotions might be detected differently across different sensing channels.

Session 2: Driving Situations & Resulting Emotions

In the next session we broke out into groups of six and did an exercise where we tried to remember significant emotions we had experienced while driving. Each person in the group was asked to put the emotion on one post-it and the situation that inspired the emotion on a second post-it. We then had people envision the future of driving – whatever level of autonomy they felt comfortable with. We asked them to envision how they thought their emotional experience would change and to describe how their envisioned future technology brought about that change.

At the end of the session, group leaders presented out the collective findings. Traffic inspired universally negative emotions, but with a wide range of arousal levels – from exhaustion to rage. Having the road to yourself inspired universally positive emotions from peace to joy. Imagined future tech inspired both positive and negative emotions, from envisioning how it would solve problems (automation allowing you to relax) to fear of how it could go wrong and take away freedom and joy of driving.


Session 3: Measurement Methods

● Camerao Used for: facial expressions/gestures/body postureo Pro: easy to use, cost effective, pervasive, abundance of computer vision algorithms, open source algorithms (OpenCV), contact freeo Con: privacy● Heart rate & Heart rate variability○ Used For: Stress, Emotion, Health○ Pro: reliable measure for arousal;○ Cons: needs skin contact or capacitively sensed but expensive and only partially reliable● Microphones○ Speech (content and non-verbal)○ Con: Privacy● Driving performance/style● wearables● questionnaires● thermal camera● brain waves (fNIRS and EEG)● blood pressure (-no easy way of measurement)● breathing● gesture (- angle of view)● personal driver profile● head position● Physiological (EDA, EMG, EOG, ECG)● Steering wheel● Breathing rate and depth (good indicator of sleepiness)● Pheromones - > blood sugar level● Chemical sensors● can-bus (vehicle state)● human in the development cycle → self-reported metrics → seek empathy● pressure sensors on seats● eyetracking: pupil dilation, eye gaze (attention)● causal analysis● microsteering● chemical sensors, e.g. blood sugar level● choice of clothing● calendar/whatsapp/smileys● historical data● interpersonal communication settings● body count detector● hobbies/sports performance● road conditions (potholes, …)● evolution of emotional states over timeDo you read all of them? which are suitable for driving? they should be personalized!

Session 4: Rapid Prototyping!

Prototype 1

The idea here is to calm down the driver during some negative affect episodes. The system will change the colour of the cabin in order to induce the driver in a more positive emotion, will diffuse some nice fragrance (e.g., lavender) and will propose the driver to listen to his/her favourite music with the special Empathetic DJ application!

Prototype 2

Idea of the prototype was to develop an intelligent adaptive speech-based HMI that is able to include the driver’s personality within its attempts to regulate the driver’s / car user’s affective state caused by external events, e.g., somebody blocks your way. Adaptivity reaches its peak with the HMI’s ability to convey ironic comments to regulate driver’s / car user’s anger.

Prototype 3:

Automatic Child Separation System

Our prototype was a child isolation system that activated when it detected that children in the back seat were fighting. The system consisted of a system of “separator” panels that activated on acoustically detected yelling or visually detected poking, hitting, or “space encroachment” by one child on another. The separators lowered after the children calmed down both vocally and physically. Other variations were to have screen time mediated by quiet (more quiet, more screen time)

Prototype 4:

Situation: I’m on my way to an interview with Very Important Company. Suddenly, a traffic jam comes up and it becomes clear that I’ll be too late. The carsharing-fleet operator sends a message through my personal in-car assistant asking whether he should inform Very Important Company that I will be too late. Additionally, the assistant offers to take responsibility for the delay because of miscalculated optimal routes. I accept the offer and therefore the message is sent to the company. Subsequently, the assistant asks Very Important Company whether a skype connection should be set up. They answer with yes and therefore the interview is then started via skype while the passenger is in the traffic jam.

Prototype 5

Prototype 6

Session 5: Challenges and Outlook

Session 6: 1-minute Wrap-Up Videos

Video 1:● emotion detection is less of a problem because it could even be input by the user directly● ‘the issue is really how to design a scheme that uses the control that is available to the car to counterbalance this (the emotion) and how to tailor this for an individual person’Video 2 - (not sure of I get his message right):● his work is on emotion/state detection● as soon as that can be done, can we do the interfence of causality across all vehicles or can we customize them per user?Video 3:● ‘how do I detect the cause of an emotion from a detected emotion state?’Video 4:● personalize emotion recognitionVideo 5:● which group of measurements are good to differentiate between valence and arousal levels?● has there been research on the acceptability of such systems?Video 6:● measurements need to be multimodal to account for the fact that there are different types of users and measurement uncertainties● use odor as countermeasure● which methods are most effective as countermeasures and when to use which ones● depending on the person, the invasiveness needs to be adjustedVideo 7:● you always have a degree of uncertainty on the measure of emotion ● each system needs to be trained individually for each user● countermeasures need to be subtleVideo 8:● especially emotions induced by vehicle behavior are interesting● depending on the situation, the car should either seek the human’s help or try to comfort himVideo 9:● which emotions are important in the automative context?● how can these emotions be measured, espcially in the car?● will these measurement systems be accepted?
Mein Film1.mp4
Mein Film2.mp4
Mein Film3.mp4
Mein Film4.mp4
Mein Film5.mp4
Mein Film6.mp4
Mein Film7.mp4
Mein Film8.mp4
Mein Film9.mp4

Group picture!