Flock-follow Experiment

Live at https://crcdng.github.io/flock-follow/.

This is an experiment I learned in a workshop some time ago (I don't quite remember the exact occasion):

Ask a group of people to walk around in a restricted space, so that flocking behaviour can emerge. Let them keep moving while trying not to bump into each other.

Observe the behaviour for a while.

Then introduce the "follow-rule". From a certain moment, signalled by a clap, each person:

1. secretly selects another random person, and
2. tries to keep a constant distance to that person.

Observe the behaviour again. What has changed?

I have done the experiment a number of times with groups of students. Afterwards I have shown this simulation, which attempts to model the phenomenon.

This code grew from a rewrite of the p5.js flocking example by Daniel Shiffman from "The Nature of Code" (http://natureofcode.com).

To run locally, unzip or clone this repository, then start a webserver, e.g. live-server in the main directory.

DONE

• ES6
• implement parameter UI with Materialize
• implement "follow" logic
• expose neighborhood parameters
• save/load parameters
• implement run/pause
• implement variable number of boids
• implement "highlight one" (klick on one boid while paused to highlight it)
• implement tracing of selected boids
• lint the code with semistandard

IDEAS

• implement a small random error in the following distance

• display graphs of parameters over time, e.g. global cohesion

• save trace data

• add an option to show a video of the experiment with humans and the simulation side-by-side

• implement more behaviours from Reynolds' paper

• show parameter values in the UI

• simulation speed

• better handle "load parameters" (path)

• select colors for boids / background

• modularize the code, make p5.js work with ES6 modules

• use function parameters instead of global variables

• optimize (avoid vector creation in the draw loop)

• optimize (reduce multiple runs through the boids array in flocking behaviors into one)

• add sonification

• extend the human experiment with switching followees, alternating between following and non-following mode, varying speed, different group and room sizes.

References

p5.js

https://p5js.org/examples/simulate-flocking.html

http://natureofcode.com/book/chapter-6-autonomous-agents/

Reynolds, Craig W. “Flocks, Herds and Schools: A Distributed Behavioral Model.” ACM Siggraph Computer Graphics 21, no. 4 (1987): 25–34.

Reynolds, Craig W. “Steering Behaviors for Autonomous Characters.” In Game Developers Conference, 1999:763–782, 1999.