Week 1: Conceptualisation

 

What I worked on:

 

I used the first week scoping the core concept for my graduation project which is a interactive physical computing installation using a Devil Ivy plant connected to an Arduino and an external display running custom visual animations. I sat there and read texts around plant neurobiology, focusing primarily on Monica Gagliano’s research into plant cognitive ecology and bioacoustics. I also began exploring the broader Plant Intelligence movement to ground the work within non-human agency and critical design.

 

Decisions made and why:

 

I decided to frame this as an artistic, practice-based project rather than a purely technical engineering exercise because I feel like merging art and code is my strength. Rather than treating the plant simply as an ornamental sensor casing, my goal is to investigate how visual graphics can mediate interspecies communication and challenge human “plant blindness”. I chose the Devil Ivy because its hardiness makes it resilient for sustained interaction and capacitive sensing.

 

Problems:

 

I initially considered attempting to build a complex 3D sim of the plant in Unity or Unreal Engine but that’s not my expertise and also it shifts the focus away from the living organism’s physical presence and tactile reality.

 

Next Steps:

 

I will map out the sensor architecture and benchmark Qt/QML against my hardware constraints.

 

 

Week 2: Technical Arch and Prototyping

 

What I worked on:

 

This week I focused on defining the technical pipeline connecting the hardware and graphical interface. I mapped out the communication flow; the Arduino reads environmental and capacitive touch values from the plant, passes data into serial packets over USB and transmits them to the host machine (MacBook during development, transitioning to Raspberry Pi for standalone exhibition). I also began mocking up asset concepts in Adobe Illustrator for plant state transitions.

 

Decisions made and why:

 

 I decided to build a frontend container using Qt Creator and QML rather than using standard web tools because QML handles lightweight, hardware-accelerated UI rendering efficiently on low-power devices like the raspberry Pi. For the animations, I decided to export After Effects vectors via Lottie/JSON instead of heavy video renders to keep memory overhead low.

 

Problems:

 

Using basic resistive moisture pins highlighted long-term corrosion risks in wet soil. I need to get a capacitive sensor to prevent soil contamination and sensor decay over an extended exhibition

 

Next steps:

 

Acquire a capacitive touch breakout board to touch sensitivity through the leaves and write a C++ serial reader in Qt.