Inverse Kinematics
- Introduced by
- Pieper, D. L. et al.
- Year
- 1968
- Type
- Problem-solving method for robot control
- Family
- Numerical computation
What does it do?
Inverse kinematics is the computational problem of determining the joint angles required to position and orient the end-effector (tool) of an articulated mechanism at a desired pose (position and orientation).
In contrast to forward kinematics, which computes end-effector position from joint angles, inverse kinematics solves the reverse mapping. This is essential for robot control: given a desired target location, IK finds the joint commands that achieve it.
The intuition behind the method
The core challenge is that while forward kinematics is a straightforward (usually unique) transformation, inverse kinematics is often non-unique or has no closed-form solution. For a 6-degree-of-freedom robot, multiple joint configurations may reach the same end-effector pose. Some robot geometries have closed-form solutions; others require iterative numerical methods. The problem becomes more difficult when the workspace is limited, singularities exist, or joint constraints (limits, collisions) must be respected.
History
Inverse kinematics emerged as a critical problem in robotic control during the 1960s-1970s. Pieper's 1968 dissertation addressed the IK problem and identified conditions for closed-form solutions. Over decades, researchers developed analytical methods for standard robot architectures and iterative numerical methods (Jacobian, gradient-based) for general cases. Modern robotics libraries incorporate sophisticated IK solvers combining both approaches.
When to use it
Use inverse kinematics whenever you need to program a robot to reach a specified target location. Essential for trajectory planning, pick-and-place operations, and any task-space control application. Assume the target pose is reachable within the robot's workspace; if not, fall back to the closest reachable pose.
Limitations and considerations
- Closed-form solutions exist only for specific robot geometries (e.g., 6-DOF robots with three consecutive revolute joints)
- Multiple solutions (multiple joint configurations reaching the same pose) complicate selection
- Singularities (configurations where the end-effector loses degrees of freedom) cause numerical instability
- Iterative methods are slower and may not converge; they require good initial guesses
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- Enables intuitive task-space programming (specify desired end-effector position)
- Closed-form solutions are fast and deterministic, suitable for real-time control
- Numerical methods generalize to any robot geometry and are numerically robust
- Allows incorporation of constraints such as joint limits and collision avoidance
Applications
- Industrial robot programming: specifying pick-and-place locations for assembly robots
- Path planning: computing joint trajectories that move the end-effector along a desired path
- Teleoperation: converting human operator commands in task-space to robot joint commands
- Motion capture and animation: converting captured end-effector poses to joint angles for playback or animation
How it works
How it works
Define the desired end-effector pose (position and orientation), the robot's kinematic structure (link lengths, joint types), and any constraints (joint limits, collision avoidance).
Choose an appropriate IK method based on the robot structure: analytical (closed-form) for simple geometries, or numerical iterative methods for general cases.
For analytical IK, derive closed-form equations from the kinematic equations. For numerical methods, set up the optimization or iteration framework.
Solve the IK equations using the chosen method: direct algebra for 3-link planar arms, Jacobian-based iteration (Newton-Raphson) for general robots, or gradient descent for constrained problems.
Return the joint angles, check for validity (within joint limits, collision-free), and select the best solution if multiple exist.
Common pitfalls
- Attempting closed-form IK for a general robot geometry, then giving up when the algebra becomes intractable
- Failing to validate that the target pose is actually reachable before solving
- Selecting an IK solution that causes joint discontinuities or high speeds in a trajectory
Frequently asked questions
Why does my robot have multiple solutions to reach the same target?
Redundancy (more than 6 DOF) or specific robot geometries can create multiple configurations with the same end-effector pose. This is advantageous: you can choose the solution that avoids obstacles or joint limits. Use the null space to adjust joint angles without changing end-effector position.
What is a singularity and why does it cause problems?
A singularity is a configuration where the robot loses one or more degrees of freedom (e.g., at full arm extension or when joints align). At singularities, the Jacobian is singular, causing infinite joint velocities for finite end-effector velocities. Avoid trajectories passing through singularities.
Should I use analytical or numerical IK?
Use analytical (closed-form) IK if your robot geometry admits a solution; it is fast and deterministic. Use numerical methods for complex geometries or when constraints complicate the problem. Hybrid approaches (analytical for reachable poses, numerical fallback) are common in practice.
How do I handle joint limits in IK?
If an IK solution violates joint limits, use a constrained optimization approach: minimize the distance to the target while keeping all joint angles within their limits. This produces the best-effort solution if the exact target is infeasible.
Related methods
Sources
Craig, J. J. (2005). Introduction to Robotics: Mechanics and Control (3rd ed.). Pearson Education.
ISBN: 0-13-123629-6Copied.Spong, M. W., Hutchinson, S., & Vidyasagar, M. (2006). Robot Modeling and Control. John Wiley & Sons.
ISBN: 0-471-64990-2Copied.Pieper, D. L. (1968). The kinematics of manipulators under computer control. Ph.D. Dissertation, Stanford University.
Open publication referenceCopied.