Swarm
Swarm coordination studies decentralized policies for formation, flocking, consensus, and area coverage. The chapter emphasizes local rules, communication assumptions, and emergent global behavior.
One path, six algorithms
Every simulation in this chapter steers the same reference: a figure-8 at swarm scale, swarm_figure_8_ref. They previously each invented their own motion — two circular orbits, a fixed corner goal, and two that never translated at all — so comparing them meant first accounting for the fact that they were not doing the same thing.
A figure-8 rather than an orbit because a circle lets a formation settle into one steady bank and hold it forever, and so never shows whether the group can be pulled through a reversal and put back together. The crossing is the part worth watching.
What each algorithm attaches to the guide point differs — a leader, a virtual structure's origin, a potential well, a coverage region — but the path they trace does not.
Core Questions
- What information must be shared for stable formation behavior?
- How do local interaction rules scale with team size?
- Which methods retain robustness under dropouts and disturbances?
Algorithms
Prerequisites
- Graph theory basics for multi-agent systems
- Consensus and stability concepts
- Potential fields and geometric formation models