State machines

Control & autonomyIntermediate30 minBuild a sumo state machine ↓Before this: Conditional logic (if / else)

What you’ll learn

What a state machine is
How to draw one and turn it into code
Why it beats a pile of nested ifs

States and transitions

A robot is always in one state: SEARCH, ATTACK, ESCAPE. Each state has its own simple behaviour. Transitions are the rules for moving between states: “if I see the opponent, go from SEARCH to ATTACK”.

        sees opponent
 SEARCH ─────────────► ATTACK
   ▲  ▲                  │
   │  │ edge clear       │ sees edge
   │  └──── ESCAPE ◄─────┘
   │          ▲
   └──────────┘ (from any state: sees edge → ESCAPE)

In code

let state = 'SEARCH';

function loop() {
  const edge = !lineLeft() || !lineRight();
  if (edge) state = 'ESCAPE';           // top priority transition

  if (state === 'SEARCH') {
    motors(30, -30);
    if (distance() < 180) state = 'ATTACK';
  } else if (state === 'ATTACK') {
    motors(80, 80);
    if (distance() >= 180) state = 'SEARCH';
  } else if (state === 'ESCAPE') {
    motors(-60, -20);
    if (!edge) state = 'SEARCH';
  }
}

Why it helps

Each state is short and easy to understand.
You can print the state to see exactly what the robot thinks it’s doing.
Adding a behaviour means adding a state, not untangling ifs.

Example: a traffic light

State	Lights	Next state after
RED	red	5 s → GREEN
GREEN	green	4 s → YELLOW
YELLOW	yellow	1 s → RED

enum State { RED, GREEN, YELLOW };
State state = RED;
unsigned long since = 0;

void loop() {
  unsigned long t = millis() - since;
  if (state == RED && t > 5000)    { state = GREEN;  since = millis(); }
  if (state == GREEN && t > 4000)  { state = YELLOW; since = millis(); }
  if (state == YELLOW && t > 1000) { state = RED;    since = millis(); }
  // set the LEDs for the current state here
}

Example: a delivery robot

FOLLOW_LINE → (junction) → TURN → (line found) → FOLLOW_LINE → (end marker) → DROP → (done) → RETURN.

Finished this lesson?