Ready Set Mow
Design
Development
3D Art
Overview
role
Sole designer, developer, and artist
tech
C#
Skeletal tracking
Computer vision
tools
Unity
Blender
mediapipe
Photoshop
Ready, Set, Mow! is an alternative arcade racing game where the controller is an actual zero-radius riding lawn mower. You use the lawn mower to control your vehicle in the game as you race to see who can mow the fastest in your neighborhood.
I made this project for Maker faire Orlando – a family-friendly showcase of invention, creativity, and resourcefulness. It utilizes computer vision to integrate the controls of the zero-radius lawn mower into a unity game!
My Role and Responsibilities
Hardware integration
As the sole developer for this project, I had to overcome the challenges of integrating a working zero-radius lawn mower into a video game. I explored a number of solutions, but ultimately settled on utilizing computer vision for hand tracking, calculating the input values from the position of the hands within view, and then sending that data to the unity instance running my game.
Game development
On the unity side of the project, the primary challenge to making this game was getting the visuals right for the grass clippings that you mow in the time I had. I ended up implementing a custom compute shader to instantiate and render all the different grass blades effeciently. Then, I utilized a special buffer for storing grass visibility that I could dynamically change at runtime when detecting collisions.
Lessons learned
I had a blast making this project! The key takeaway I had here was related to speed : I only had around three weeks worth of time to get this project together, and the biggest risk area was getting the lawn mower integrated with the game. While considering options, I was skeptical that a hand tracking solution would be capable of reaching the fidelity I needed to be playable. However, my other options included a hall effect sensor or a sliding potentiometer, both of which would have required time developing custom hardware that would fit over the handles of the mower.
Because I was able to implement a minimum viable solution so quickly, I was then able to polish the fidelity of the tracking in iterations, changing the webcam images contrast, resolution, and color tone to find the most reliable way to track the hands. Ultimately, this was the right way to go and I am proud of the result.
