Reactive Robots: Braitenberg Vehicles

In 1984 Valentino Braitenberg showed that a vehicle with two sensors wired straight to two wheels can look like it has feelings. There is no program and no map, only wiring.

Sensor gain Base speed Fear: same side, excites Aggression: crossed, excites Love: same side, slows Explorer: crossed, slows

Click or drag on the canvas to move the light.

The wiring

Each vehicle has a left sensor $s_L$ and a right sensor $s_R$ that get stronger closer to the light. With excitatory, same-side wiring, each sensor speeds up the wheel on its own side:

$$v_L = k\,s_L \qquad v_R = k\,s_R$$

Crossing the wires swaps them. Inhibitory wiring slows the wheel instead: $v = v_0 - k\,s$.

What to notice

Fear turns away from the light and speeds up when it is close, so it looks like it is running away. Aggression turns toward the light and speeds up, so it charges straight at it. Love turns toward the light but slows down, so it comes to rest gazing at it. Explorer slows near the light but turns away, so it lingers and then wanders off to find another.

This is the core of reactive robotics: tight sensor-to-motor loops with no planning. Many real robots still use reactive layers for fast reflexes, such as stopping when a bumper is pressed.