Road to RubyKaigi: Play Hard(ware)

by makicamel

Embed

Start on current slide

Slide 1

Slide 1 text

Road to RubyKaigi: Play Hard(ware) @makicamel RubyKaigi 2026 2025.04.22.

Slide 2

Slide 2 text

About Me @makicamel / Maki Kawahara Loves Ruby , beer and sake Co-Organizer of PicoPicoRuby Creator of "Road to RubyKaigi"

Slide 3

Slide 3 text

Road to RubyKaigi 　　 Road to RubyKaigi 　 https://github.com/makicamel/road_to_rubykaigi

Slide 4

Slide 4 text

Road to RubyKaigi A Ruby game by Rubyists, for Rubyists A side-scrolling action game played in the terminal Defeat bugs, escape deadlines, and make it to RubyKaigi

Slide 5

Slide 5 text

Road to RubyKaigi

Slide 6

Slide 6 text

PicoRuby

Slide 7

Slide 7 text

PicoRuby PicoRuby is very HOT at RubyKaigi 2026!

Slide 8

Slide 8 text

PicoRuby Ruby for embedded systems Runs on microcontrollers e.g. Raspberry Pi Pico 2W

Slide 9

Slide 9 text

Microcontroller Connect various I/O to build your own device

Slide 10

Slide 10 text

Acceleration

Slide 11

Slide 11 text

Acceleration What is acceleration? The rate of change of velocity e.g. A car starting to move: positive acceleration A car cruising at constant speed: acceleration is 0

Slide 12

Slide 12 text

Acceleration Acceleration Expresses changes in motion as numbers Lets us extract features and read movement 　　 3-axis accelerometer module KXR94-2050 　 https://akizukidenshi.com/catalog/g/g105153/

Slide 13

Slide 13 text

Acceleration With an accelerometer, a human walking can make the game character walk!

Slide 14

Slide 14 text

Accelerometer 3 axes x: left-right y: front-back z: up-down Reads acceleration on each axis

Slide 15

Slide 15 text

Accelerometer Waveform while standing still Almost no change in acceleration

Slide 16

Slide 16 text

Accelerometer Waveform while walking Draws periodic arcs

Slide 17

Slide 17 text

Accelerometer Find patterns of change to read "motion"

Slide 18

Slide 18 text

Walk Detection

Slide 19

Slide 19 text

Walk Detection: Time Acceleration at a single moment A single point alone can't reveal "motion"

Slide 20

Slide 20 text

Walk Detection: Time Acceleration change over a time window "Moving" means changing over time Observe variation within the window

Slide 21

Slide 21 text

Walk Detection: Motion Intensity Amount of acceleration change in a time window = motion intensity How scattered the acceleration is

Slide 22

Slide 22 text

Walk Detection: Motion Intensity Condense acceleration change over a time window into a single number Represents motion intensity Still / walking become distinguishable Waveform information is lost Periodicity, rhythm of the wave Acceptable for this case If motion intensity exceeds a threshold, the player is "walking"

Slide 23

Slide 23 text

Walk Detection: Extracting Motion Intensity

Slide 24

Slide 24 text

Walk Detection: Extracting Motion Intensity 1. Extract motion intensity per axis 2. Sum the extracted values

Slide 25

Slide 25 text

Walk Detection: Per-Axis Motion Intensity e.g. Z axis

Slide 26

Slide 26 text

Walk Detection: Per-Axis Motion Intensity Mean mean = samples.sum / samples.size Deviation from the mean for each sample dev(n) = sample[n] - mean Squared deviation dev(n)² Mean of squared deviations (dev(1)² + dev(2)² + ... dev(n)²) / samples.size

Slide 27

Slide 27 text

Slide 28

Slide 28 text

Slide 29

Slide 29 text

Slide 30

Slide 30 text

Walk Detection: Per-Axis Motion Intensity Motion intensity along the Z axis over the window

Slide 31

Slide 31 text

Walk Detection: Per-Axis Motion Intensity Motion intensity along X, Y, Z axes over the window

Slide 32

Slide 32 text

Walk Detection: Per-Axis Motion Intensity Sum the per-axis motion intensities def motion_intensity Math.sqrt(var(0) + var(1) + var(2)) end Motion intensity over the window

Slide 33

Slide 33 text

Walk Detection: Per-Axis Motion Intensity Walking starts when motion intensity exceeds the threshold class SignalInterpreter def walk_started? @window.motion_intensity > Config.start_threshold end

Slide 34

Slide 34 text

Walk Detection After this, add stop detection, running behavior, jump detection, and so on, done!