humanoidsdata.com

Search

Search datasets, articles, and glossary terms for humanoids and embodied AI.

← All glossary terms

Hardware & control

Robot kinematics

Robot kinematics describes the geometric relationship between a robot’s joint configuration and the position, orientation and velocity of its links or end-effector, without modelling the forces that cause the motion. Forward kinematics computes pose from joint values; inverse kinematics searches for joint values that achieve a requested pose.

Also known as: kinematics, robotic kinematics

Updated

Forward and inverse kinematics

Forward kinematics evaluates a geometric model of the robot. Given joint angles or displacements, it returns the pose of an end-effector or another link in a declared reference frame.

Inverse kinematics asks the opposite question: which joint configuration produces a requested pose? A target may have no solution, one solution or several solutions. Numerical solvers therefore minimise pose error while enforcing joint limits, collision constraints and other feasibility checks.

The Jacobian and differential motion

Velocity kinematics uses a configuration-dependent Jacobian matrix to relate joint velocity to end-effector velocity. Modern Robotics Chapter 5 also shows how loss of Jacobian rank identifies kinematic singularities, where motion in one or more task-space directions becomes unavailable.

Differential inverse kinematics uses that local relationship to calculate joint velocities for a desired end-effector velocity. A pseudoinverse gives a least-squares solution, but practical controllers often add damping, joint-limit handling and collision constraints.

Why kinematic metadata matters

Kinematic quantities are meaningful only with the robot model and coordinate frames that generated them. A dataset should state joint order, angle or distance units, base frame, end-effector frame, calibration and whether a pose was measured, estimated or computed from forward kinematics.

Kinematics alone does not predict inertia, force, friction, compliance or contact response. Those belong to dynamics and control, even though a deployed controller commonly uses both kinematic and dynamic information.

Sources