Humanoids and Legged Robots
Do Humanoid Robots Need Legs? The Case for Wheels
Legs cost energy, complexity and safety margin. They earn their place only where stairs, thresholds or clutter cannot be removed, which is a narrower set of sites than it sounds.
Legs buy access to stairs, kerbs and cluttered floors, and they cost energy, complexity and a fall risk that no wheeled machine carries. In a flat, purpose-built facility a wheeled mobile manipulator does the same transfer work at a fraction of the standing power and with a mature safety framework behind it. Legs are worth it where the environment cannot be changed, and that is the whole argument.
The honest ledger
| Criterion | Legged | Wheeled |
|---|---|---|
| Standing power | 100 to 400 W | 20 to 60 W |
| Actuated joints in the base | 12 to 14 | 2 to 4 |
| Stairs and kerbs | Yes | No, without a lift or ramp |
| Payload limit driver | Balance and joint torque | Tipping stability, usually higher |
| Fall risk | Present, and a hazard in itself | Effectively absent |
| Safety standard | None specific | ISO 3691-4, mature |
| Runtime | 2 to 5 h | 6 to 12 h |
| Cost of the base | High | Moderate |
| Maintenance points | Many, high load | Few, low load |
The runtime row compounds the power row. A wheeled base with the same battery does more work per charge because it does not pay to stand, and standing is 20 % to 40 % of a typical duty cycle.
The middle ground is winning
Several designs deliberately split the difference, and in practice they cover most of the tasks that motivated humanoids in the first place.
- Wheeled base with a humanoid torso. Two arms, a human-height reach envelope and a stable wheeled platform. It handles workstations built for people without paying for balance.
- Wheel-leg hybrids. Powered wheels at the end of articulated legs, which roll efficiently on flat ground and step over an obstacle when needed.
- Tracked bases. Handle thresholds and rough ground without the balance problem, at the cost of manoeuvrability and floor wear.
- Lift columns. A wheeled base with a vertical axis reaches from floor level to a high shelf, which is one of the main reasons a human-height machine was wanted.
When legs genuinely win
- Stairs that cannot be avoided. Older buildings, ship interiors, plant structures with grating and steps.
- Cluttered or uneven ground. Construction sites, disaster response, agricultural terrain.
- Narrow footprint in tight spaces. A biped occupies less floor area than a stable wheeled base of the same reach.
- Environments that must not be modified. Heritage sites, customer premises, temporary deployments where installing ramps is not an option.
- Tasks needing a human-like workspace envelope, such as reaching into a car footwell or working at a bench designed for a standing person.
Point four is the strongest commercial argument. The cost of modifying a facility for wheeled robots is real and sometimes prohibitive, and a legged machine that walks into an unmodified building removes an entire capital project. Where a facility is being built new, that argument disappears, and it is why greenfield sites almost never specify legs.
A three-question test before specifying legs
Run this before any humanoid procurement. If all three answers are no, the requirement is for a mobile manipulator and the humanoid form is buying nothing.
| Question | If yes | If no |
|---|---|---|
| Does the route include steps, kerbs or thresholds above 30 mm that cannot be ramped? | Legs or a wheel-leg hybrid | Wheels |
| Is the floor ever obstructed by cable, pallet debris or uneven surface that a 100 mm wheel cannot cross? | Legs or tracks | Wheels |
| Is modifying the building forbidden or more expensive than the robot? | Legs | Wheels plus modification |
The 30 mm figure is the practical threshold at which a typical 150 mm to 200 mm drive wheel starts to struggle under load, and it is worth measuring rather than estimating. Many sites that feel impassable turn out to have three specific thresholds, each of which can be ramped for a few hundred currency units.
Frequently asked questions
Are legs necessary for a humanoid robot?
Only where the environment cannot be changed. On flat purpose-built floors a wheeled base performs the same transfer and manipulation work with lower power, longer runtime, higher payload and a mature safety framework.
How much more power do legs use?
Standing alone costs 100 W to 400 W on a legged machine against 20 W to 60 W on a wheeled base, and standing is typically 20 % to 40 % of a duty cycle. Locomotion is also less efficient over flat ground.
What is a wheel-leg hybrid?
A base with powered wheels mounted at the end of articulated legs. It rolls efficiently on flat floors and steps or lifts over an obstacle when needed, capturing much of the access advantage without paying the full balance cost.
Is a falling robot a real risk?
Yes, and it is a hazard class that contact-force standards do not address. A 65 kg machine toppling is an impact event, which is why legged deployments have so far relied on site-specific safety inspection rather than product certification.
Why do greenfield sites rarely choose legs?
Because a new facility can be built flat, with ramps and lifts, so the access advantage of legs disappears while the energy, cost and safety penalties remain.
Sources
- ISO 3691-4, safety requirements for driverless industrial trucksInternational Organization for Standardization, the framework covering wheeled bases
- ISO 13482, safety requirements for personal care robotsInternational Organization for Standardization, includes provisions on fall hazards for legged robots
- Unitree G1 entry in the ROBOTS guideIEEE Spectrum, published runtime and mass figures used for the power comparison