Welcome to our page!

Brief:

Creating trustworthy AI experiences for the users, i.e. passengers and drivers of Ford vehicles.

Our take

Driving is often a solo activity. Even in a car full of passengers, the driver is single in his activity of driving, often leaving him unaccompanied, resulting in a range of emotions. From being stressed and fatigued from a long day, to being energetic and calm, to feeling bored.

To make the AI feel more trustworthy, the AI should feel like a good partner and friend for the driver.

Who is a Good Friend?

Concepts

We want to create a good friend for the driver, how can we achieve this goal? We found that most of the driving experience is functional, and can be carried out in a safe way, preventing accidents and connect to their multiple devices.

But there is also another aspect to consider: emotions! An emotion determines the mood of the driver and the mood is like the background color of what we do. We identify 4 main moods based on the Pick-a-mood by Martijn Vastenburg, Natalia Romero,? Pieter Desmet.

From this 4 moods, we can identify more specific moods that can be recognized and stimulated to improve the driving experience.

What can allow us to create a good friend to accompany the single driver is identifying the emotion, train a model to understand the emotion and provide a reaction, either to enhance the current emotion, or slightly change the scenario to enhance a different emotion.

1. Understanding

Prototype 1 and concept 1 ? Car music assistant

Scenario- Your car understands your non-verbal communication cues, and your car could help you do some easy tasks.

Concept- Using your facial expressions and hand gestures, the AI recognises what to do and controls the music accordingly. Suggesting appropriate music, to controlling the volume of the music.

https://youtu.be/eLOoB7vAxy0

2. Fatigue

Prototype 2 - Wake Up!

Scenario- You?re the designated driver, you?re coming back from attending a music festival with your friends, and you have to drive back home for a few hours. Your friends have fallen asleep in the back seat. You are extremely tired. The AI detects this, and offers to make physical changes that would help you feel more awake.

Concept- The car AI uses facial recognition and understands that you?re beginning to feel sleepy. It checks up on you and if you tell it you?re tired then -It knows what you like doing in such situations, so it offers to open the windows, play music, turn on the AC for you. You can select one and it does it for you.

3. Cheer Up!

Prototype 3- Trivia Time!

Scenario- You?re driving to work, and you are feeling a little stressed.The AI detects your mood and it tries to cheer you up! It engages you in an entertaining trivia game.

Concept- This was inspired by the previous concept. It is a trivia game that works

Prototype 4- Car-a-oke

Scenario- You?re driving and listening to the radio, a familiar song comes up and you start humming the tune and singing to it, The ai detects it, and encourages you to sing louder with it, by giving you feedback in the form of changing ambient lights.

Concept- The concept is using light colour to reflect and conncet the emotions from the drivers. The program will analyse the spectrum of your sound and change the light through your the changing of your voice.

Also, for futher develoopment another issue will influenc the light colour is your face expressions, by applying facemesh to machine learning, we can detect the emotions by the face expressions of drives, in this way, the emotions is connected to our prototype deeper.

Future steps

• Focus in a specific target group (professional driver rather than particular driver)

• Explore more ways to recognize the mood of the driver (face recognition, voice recognition, grip on the steering wheel)

• Define what reactions can provide the effect that we are looking for (we defined just light and voice as predetermined reactions)

• Exploring the other sensory of feedback (not just voice and visuals), using Arduino and sensors.