Drivatar
Duration
Type
Team
My Role
4 Weeks
AI Audio Application Design
Hannah Meng, Siddharth Sudhakaran, Chen Tseng, Hairong Long, Ke Shen
Research, UIUX Design
Overview
Innovating Audio Dynamics
Imagine a world where the smart speaker, powered by generative AI, tailors unique audio experiences to individual preferences and situations.
We delve into creating diverse audio landscapes, overcoming the challenge of refining inputs for comprehensive and consistent audio narratives.
The mission is to design universally appealing AI compositions, transcending boundaries and resonating with listeners from diverse backgrounds. The goal is to redefine audio dynamics with cutting-edge technology and nuanced perceptions.
Client Interview
The following are the main three things we aim to discover based on the client’s expectation of this project.
Nuance
Able to capture the users’ subtle facial expressions
Real Time
Deliver the suitable solutions at the moment
Simple
Able to understand the user’s preference and require minimum actions
Process
Research
68%
People from 18-34 listen to music everyday
70%
Of them drive the most kilometres among other age groups
Music does:
-
Reduce depression and anxiety
-
Trigger a specific emotional response
-
Beneficial effect on driving style and safety-related performance
66%
Of them experience anxiety while driving
Competitors
AI Generated Music
-
Complicated UIUX design
-
Limited Contextual Adaptation
-
Focused on General Compositions
Music Apps
-
Manual Input Dependency
-
Limited Personalization
-
Lack of Emotion-Driven Adaptation
Meditation Apps
-
Manual Input Dependency
-
Limited Personalization
OPPORTUNITIES
Accessibility
Ensure Drivatar is user-friendly for everyone
Multifaceted Inputs
Allow diverse input methods, including click, voice, type, for instructions
Adaptive Algorithm
Implement an advanced algorithm for continual learning and precise music recommendations without explicit commands
Goal
Our proposed solution, Drivatar, is an advanced in-car music system that intuitively adjusts the playlist in real-time based on the driver’s emotional state, vehicle data, and current traffic conditions.
By integrating emotion detection technology and real-time traffic data analysis, the system curates music that aids in reducing stress and enhancing focus, thereby promoting safer driving.
This hands-free, AI-driven approach ensures that the driver remains undistracted, making the driving experience both safer and more enjoyable.
User Stories
User stories have been pivotal in Drivatar's development, shaping user-centric features that align seamlessly with expectations. These narratives, capturing diverse perspectives, guide everything from emotion detection to playlist customization, ensuring a direct connection to user experiences.
As A...
01/
Daily Commuter
I Want...
A product that allows me to enjoy good music while stuck in traffic
So I Can...
Stay calm, prevent frustration, and avoid becoming a road rager
02/
Uber Driver
03/
Busy Businessman
A reliable music player that tailors music to my mood during trips
A product that I can easily skip a song through voice or simple clip
Enhance my overall sense of calmness, providing my passengers with a smoother ride
I can enjoy a seamless experience and drive safely
Key Features
The following are the features of the minimum viable product within four weeks.
Emotion Detection
Drivatar uses in-car cameras for real-time emotion detection, enhancing driving experiences and safety by accurately assessing the driver's emotional state.
Vehicle Speed Detection
The system analyzes vehicle speed using real-time data, identifying changes and traffic congestion. Simultaneously, it evaluates the driver's facial expressions to curate a playlist with music that matches the driving conditions and soothes the driver.
Recommendation System
Drivatar's AI recommendation system employs a hybrid collaborative filtering algorithm, integrating user data and individual metrics to intelligently curate personalized music selections, aiming to elevate the overall driving experience.
Song Skip
Drivers can easily skip songs using the steering wheel or cellphone, contributing to the system's AI learning algorithm, refining its understanding of the user's music preferences over time.
Wireframe
Design System
3 Steps of Onboarding
Provide users a quick introduction to Drivatar with the core features.
2 steps of Creating Account
Users effortlessly create accounts, enhancing personalization by linking to their music apps.
Refresh Playlist
The playlist adapts to changes in facial expressions or car speed detected by the in-car system.
User Test
The Drivatar user test aims to understand user playlist preferences with a driving simulator—whether they prefer discovering new songs or sticking to their playlist, assess the impact of UI color stream on driving focus, and determine user preference between our AI-curated playlist and a randomly generated one.
UI Iteration
Initially, we displayed the user's emotion with text on the screen. Following client discussion, we opted for a color bar visualization.
However, user testing revealed that this feature was more distracting than enhancing. Consequently, we removed it in the final design for a streamlined user experience.
Reflection
MVP Focus
In the four-week development window, we employed the bullseye method to identify key features for the minimum viable product (MVP). Throughout, our team prioritized making these features as comprehensive as possible in alignment with client expectations.
Assumption Verification
User tests were instrumental in clarifying user preferences, specifically their tendency towards discovering new songs or sticking to playlists, and evaluating if the UI color stream posed distractions during driving.
Cross-Disciplinary Communication
Given the project's technical emphasis, seamless communication with our developers was crucial to assess feature implementability. The collection of data significantly impacted user flow, emphasizing the importance of effective communication across disciplines.