Skip to content

Digital Prototyping for Design

Module 4

Sensing the body for meaningful interactions

Lina Bautista & Citlali Hernández

23-24 April 2024

Task: Prototype your wearable and experiment with your training model.

by Núria, Marius & Oliver

Using our newfound knowledge, we set about training a new model on the pressure inputs from a soft sensor we quickly built using conductive textiles, crocodile clips, and a Barduino board. We added a nifty feature that changed the color of a Neopixel LED mounted on the Barduino board from green to red along a gradient as the pressure increased, helping us visualize if the sensor was working correctly.

Once we had the model trained, we tested it with the sensor and it worked! We successfully sent data over OSC from the sensor connected to the Barduino to the Wekinator program, and even through to a program called Processing for further visualization and complex output.

Result

Click here to access the fabrication files.

Reflection: The four F’s of active reviewing

Facts:

Feelings:

Findings:

Future:

Module 5

Extended bodies with expressive Data

Lina Bautista & Citlali Hernández

29-30 April 2024

Task: Prototype your wearable and experiment with your training model.

by Núria, Marius, Oliver & Carlotta

This digital prototyping session centered on utilizing either our custom device or another form of movement tracking to generate artistic outputs. We explored how our daily movements could be translated into sound. Imagine if your music was influenced by your walking speed.

Training the Model

From Movement to Data

The initial step involved capturing data from the phone, such as room position, GPS, and speed, and sending it to Wekinator. Once in Wekinator, we trained our model to differentiate between walking (0) and running (1).

In a previous module, Citlali introduced us to a mobile app called “ZIG Sim,” which allows access to various phone sensors, including the accelerometer, gyroscope, microphone, and camera. This app can send sensor data via OSC messaging (among other methods) to another device for processing.

Using this, we had Mars run and walk around the room with the phone in his pocket. However, we quickly realized the room was too small to collect accurate data, so we moved the training to the roof for a larger, straighter space to gather more precise data. Although we wanted to investigate more types of movement, time constraints limited our focus to this aspect.

Steps Taken

The process from training to the final output was straightforward but required navigating through various software and interactions. Once the basics were understood, numerous opportunities emerged. The project’s goal was to explore: “What if the songs you listened to changed based on whether you were running or not?”

Grabbing Data from Phone using ZIG Sim

  1. Data Collection with ZIG Sim:
  2. We used the ZIG Sim app on a phone to collect various data points, focusing on GPS, acceleration, and placement in the Z-axis.
  3. The data was then sent to Wekinator.
  4. To account for differences in individual movement, we trained the model on multiple people running and walking.

Training the Model in Wekinator

  1. Model Training:
  2. While collecting data from the phone, we simultaneously sent it to the Wekinator application on a laptop.
  3. In Wekinator, we trained the model to recognize that running corresponds to a signal of 1 and walking corresponds to a signal of 0.

Sending Data to Max8

  1. Connecting to Max8:
  2. After completing the training in Wekinator, we connected the model to Max8.
  3. The model was configured to send a signal of either 0 or 1 based on the movement data.

Converting Received Input to Sound in Max8

  1. Sound Conversion in Max8:
  2. Using the received signal, we created a simple flow of code in Max8.
  3. This code switched between two sounds based on whether the signal was 0 or 1, effectively changing the music based on whether the user was walking or running.

Result

In the end, we successfully created a model and code that enabled us to translate a person’s running or walking into a corresponding song. It was exciting to see this in action because it suggests new ways for people to interact with their own movement, potentially enhancing experiences in dancing, running, or other sports.

Click here to access the fabrication files.

Reflection: The four F’s of active reviewing

Facts:

  • Working with Oliver, Carlotta and Marius was easygoing and pleasing.

  • I discovered new apps and technologies like ZIG Sim, Wekinator or Max8

Feelings:

  • I feel like having an unrelated delivery from our project at tjis stage was a bit distracting but also satisfying, for being able to disconnect from what we are doing and do a completely different task.

Findings:

  • Sensors are not as complicated as I imagined and they have very fun and unexpected applications.

Future:

  • I don’t plan on researching more on this topic in the near future but I’ve enjoyed doing it now and I don’t dicard it in the long run.

MicroChallenge III

Santiago Fuentemilla Garriga

7-10 May 2024

Documentation

Click here to access the 3rd MicroChallenge repository.

Reflection

  • It was a satisfying week, we achieved the goals we had: have a full stack code of our ideal vision of Laia.
  • I missed the digital fabrication part, I wish we had had more time to change Laia physically and create a new prototype.
  • In terms of coding, we got to a point that we had never imagined, we learnt a lot more than we expected.
  • We had a lot of help, specially from Pietro and Chris, they were the type of teachers that give you tools, not do the thing for you. The tools and resources such as Langchain or FasAPI were extremely usefull for the moment and the future.
  • Our master project has advanced greatly thanks to the MicroChallenge III and we appreciate a lot having time to be able to focus only on it during a full week.
Last update: June 21, 2024