Humanoid robots are machines whose body layout or behavior resembles aspects of the human form, often combining sensors, motors, control software, and AI to move and work in human-designed spaces.
What makes a robot humanoid
There is no requirement that a humanoid robot look exactly like a person. The term usually refers to machines with a human-like arrangement—such as a torso, head, arms and legs—or machines designed to perform tasks using tools, doors, stairs and workspaces built around human bodies.
A wheeled robot with two arms may solve many of the same tasks as a bipedal humanoid, so body shape is an engineering choice rather than a measure of intelligence.
Sensors, actuators and control
Cameras, depth sensors, inertial measurement units, force sensors and joint encoders help a robot estimate its surroundings and body position. Electric or hydraulic actuators move joints. Real-time controllers coordinate those joints so the robot can balance, walk, grasp and recover from disturbances.
Planning software decides how to reach a goal while respecting constraints such as joint limits, collision avoidance and stability. The control loop runs repeatedly: sense, estimate, plan, act and correct.
Where AI enters the loop
AI can help a robot recognize objects, interpret language, predict how a scene may change, select actions, or learn behaviors from demonstrations and experience. This is often called embodied AI because decisions are tied to a physical body that must deal with friction, uncertainty and safety.
Large models can make high-level instructions easier to interpret, but low-level movement still depends on specialized perception and control systems. A robot that can discuss a task fluently is not automatically able to perform it safely.
Why human-robot interaction is difficult
Robots working near people must communicate intent, maintain safe distances, respond to unexpected motion and fail predictably. NIST research on human-robot interaction studies how system performance should be measured when people and robots share tasks and environments.
Practical success depends on hardware reliability, battery life, manipulation skill, cost, maintenance and safety—not just on AI capability.
