PYTHON · NUMPY

Formation Without a Leader

Drive a group of drones into an even line using only local averaging — leaderless consensus that survives a dropout.

01Challenge

Ten drones on a line at random positions. With no central controller and no leader, each can only nudge toward the average of its neighbors. Show the group converges to an evenly-spaced formation — and that it still works if one drops out.

Predict firstA swarm of drones holds formation with no leader and no central computer — each drone reacts only to its nearest neighbors. Can that produce stable global formation?

02Model

A swarm's strength is that no single aircraft is in charge — take one out and the rest carry on. That requires decentralized coordination: each drone acts on local information only. Consensus is the simplest form: repeatedly move a fraction of the way toward your neighbors' average, and the whole group provably settles into an even line from local averaging alone, with no central controller and no global map. Because the rule is local, a dropout just removes one node; the remaining graph still converges.

03Guided practice

Worked: place drones, define neighbor averaging. Faded: iterate the consensus update, pinning the two ends. Independent: confirm even spacing, then drop a node and re-check.

04Feedback

PASS WHEN Local averaging converges the group to an even formation and re-converges after a node is removed — decentralized and fault-tolerant.

If your run fails, check:

FAIL: not converged — too few iterations or the step size a is too small.
FAIL: dropout breaks it — the rule references a fixed neighbor index instead of the current line.

05Retrieve & space

Why is leaderless better for a swarm? (No single point of failure — lose any node and the rest continue.)
What information does each drone use? (Only its neighbors' positions — purely local.)
What guarantees convergence? (Repeated averaging over a connected graph drives the group to consensus.)

06Mastery & project

Implements leaderless consensus that forms an even formation and tolerates a dropout (L3: decentralized multi-agent coordination from local rules).

This consensus rule is the coordination layer under the swarm capstone and the live WebGPU bench's formation.

GO FURTHER · THE LIVE PLAYGROUND

Coordinate a fleet in the Digital Twin lab ↗ — a live multi-domain air/land/sea fleet with a comms mesh

← Learn to Fly The Swarm Bench →

PYTHON · NUMPY

Formation Without a Leader

Run leaderless consensus to an even formation, then survive a dropout.

You read

the arrays and values already in scope

You change

the code you write in each cell

Fixed

the dataset and the checks that grade you