{wiki}

A <robot> is a programmed, actuated mechanism that performs physical tasks such as locomotion, manipulation or positioning with some degree of autonomy. Its control system belongs to the robot, alongside the mechanism. The IFR’s primary classification document reproduces this broad ISO definition and distinguishes it from the narrower definition of an industrial manipulator.<World Robotics 2025 Industrial Robots: Sources and methods by the International Federation of Robotics>[Section 1.7.1, printed p. 28, ISO 8373:2021 definitions]{ref}

A practical <robot> combines a physical mechanism, <actuators>, task instructions and, where required by the task, sensing and <feedback control>. An arm’s joint arrangement constrains the positions it can reach; a mobile mechanism must also sustain contact and balance. The degree of autonomy is a task-specific property rather than a guarantee of unrestricted reasoning or competence.

An original two-link planar arm illustrates the relationship between structure and task. For unit link lengths and joint angles $q_1,q_2$, measured with the second angle relative to the first link, <forward kinematics> gives
$$
x=\cos q_1+\cos(q_1+q_2),\qquad
y=\sin q_1+\sin(q_1+q_2).
$$
At $q_1=0$ and $q_2=\pi/2$, the endpoint is $(1,1)$; with both angles zero it is $(2,0)$. <Inverse kinematics> asks which joint values achieve a requested endpoint, while control must make the actual mechanism follow the intended motion.<Modern Robotics: Mechanics, Planning, and Control by Kevin M. Lynch and Frank C. Park>[Chapter 4, forward kinematics; Chapter 6, inverse kinematics]{ref}

Reaching a geometric target is only one part of an experiment. Position error, repeatability, applied force, execution time and behavior after a disturbance measure different aspects of performance. An arm can have an accurate kinematic model yet deviate under load; a controller can compensate only within its sensing, actuation and timing limits.

The <MIT Cheetah robot> provides a contrasting locomotion example: perception, contact planning and force generation are coordinated while the mechanism is moving. Such examples connect the physical meaning of a <robot> with <mechanics>, <geometry>, computation and experimental verification of a defined task.


 Robot (source code)