Musico
Using peripheral interaction and intelligence to create unique playlists for different contexts will help improve the music listening experience
Using peripheral interaction and intelligence to create unique playlists for different contexts will help improve the music listening experience
Musico is a context-aware speaker that uses the location, time of day and noise around the device to define the situation. Peripheral interaction with the device is used as input for the system to learn and adapt to the type of music the user listens to in a particular context.
The playlist is to be automatically selected from all playlists available to the user, based on its appropriateness to the current context. Examples of different contexts are making studying in the living room, having dinner in the dining room or cooking in the kitchen.
Studying
Having dinner
Cooking
Implicit interaction
Musico conceptually uses location, time and noise-level (decibel) to determine the context. These measures seem suitable to distinguish contexts in the home environment of the user.
Peripheral interaction
The goal is that Musico creates and updates unique playlists for every context where the user's main interaction with the speaker is in the periphery of their attention. The music can be managed by pressing, rotating and tilting.
Location
Time
Audio
Press to skip
Rotate for volume
Tilt to pause
Both the implicit and peripheral interactions are used as input for the algorithm. With the algorithm Musico can automatically generate and update the music playlists.
By default, each song starts with a rating of five stars. Skipping a song will decrease the rating by one star, whereas repeating the song will increase it.
Additionally, to playlists can be kept up-to-date by adding new songs on a weekly basis. The ratings of new songs will start with five as well.
Several prototypes were developed for the validation and evaluation of the design.
The final prototype consists of an Arduino, a rotary encoder, an RFID reader and other components to control the music. The algorithm to generate and update playlists is programmed in Processing.
User interactions included in the Musico prototype were identified through a user study, and affected the design of the speaker.
