The simplest possible unit of locomotion: a one-actuator stick-slip crawler.

The trick is that locomotion does not require multiple actuators or any joints. It requires asymmetric friction with the ground. One coiled muscle plus two asymmetric foot-pads is sufficient to ratchet forward indefinitely.

The mechanism:

• One coiled actuator ~5–10 cm long, with a small mass at each end
• At each end, a foot consisting of angled bristles or scales (think the underside of a fish, or a comb tilted backward) — high friction in one direction, low in the other
• Joule-heat the actuator → it contracts → because forward-sliding friction is low at the front foot and high at the rear, the rear end is pulled forward
• Cool → it extends → because forward-sliding friction is low at the front foot and high at the rear, the front end slides forward (the rear stays put)
• Net result: one cycle = one step forward

You can build this in an afternoon with nylon line, a drill, an oven, a battery, a 555 timer, a MOSFET, and angled paper or 3D-printed bristles. Total cost under five dollars. The whole organism has one degree of freedom and no microcontroller is strictly necessary — even a thermal bistable switch could oscillate it.

The path from there to a robotic ant:

1. Single 1-DOF crawler (above) — proves the actuator-plus-friction-asymmetry primitive
2. Bilateral two-actuator crawler — two muscles side by side, driven out of phase; differential firing produces turning
3. Tripod-gait hexapod — three coiled muscles fire together (front-right + middle-left + rear-right), then the other three. This is the classical insect gait. Six muscles, two phase groups, one oscillator. Still no microcontroller required.
4. Stigmergic swarm — multiple of the above leaving heat-trail or chemical-trail "pheromones" the next ant follows

The conceptual win for the LAL person to internalize: coiled-line actuators are doing for soft robotics what cheap servos did for hobbyist robotics in the 2000s — they collapse the cost of a degree of freedom by an order of magnitude. The right question is not "how do we control a complex robot with these" but "what's the smallest thing that walks?" — and the answer is one muscle plus two asymmetric feet. Everything else is composition.

Two things worth telling them about that aren't obvious:

• The cooling time is the rate limit. Contraction is fast (sub-second), relaxation is slow (several seconds for natural cooling on a 0.5mm fiber). Active cooling, smaller diameter, or accepting low-bandwidth motion (which ants don't need to be fast) all work. For an ant-scale crawler this is fine; ants walk at low Hz.
• Hysteresis is severe. Open-loop position control is essentially impossible. But for a stick-slip crawler that doesn't matter — you're just oscillating between two end states, not commanding a precise position. The hysteresis becomes a feature: the muscle "remembers" its last state and the friction asymmetry handles the rest.

Reference for them to read first: Haines, Lima, Li et al., "Artificial Muscles from Fishing Line and Sewing Thread," Science 343 (6173), 868–872, 2014. Then for locomotion specifically, the literature on bristlebots and "vibrobots" — same friction-asymmetry principle, different actuator.