[Paper Review] Forward and Inverse Kinematics Seamless Matching Using Jacobian

Zeeshan Bhatti, Asadullah Shah+2 more|arXiv (Cornell University)|Jan 7, 2014

Human Motion and Animation6 references3 citations

TL;DR

This paper presents a method for seamless matching between forward kinematics (FK) and inverse kinematics (IK) in 3D character animation using the pseudo-inverse of the Jacobian matrix. By computing joint rotations via the Jacobian, the approach eliminates popping or jumping artifacts during transitions, enabling smooth, automatic alignment between FK and IK systems without manual correction.

ABSTRACT

In this paper the problem of matching Forward Kinematics (FK) motion of a 3 Dimensional (3D) joint chain to the Inverse Kinematics (IK) movement and vice versa has been addressed. The problem lies at the heart of animating a 3D character having controller and manipulator based rig for animation within any 3D modeling and animation software. The seamless matching has been achieved through the use of pseudo-inverse of Jacobian Matrix. The Jacobian Matrix is used to determine the rotation values of each joint of character body part such as arms, between the inverse kinematics and forward kinematics motion. Then moving the corresponding kinematic joint system to the desired place, automatically eliminating the jumping or popping effect which would reduce the complexity of the system.

Motivation & Objective

• Address the challenge of seamless transition between forward kinematics (FK) and inverse kinematics (IK) in 3D character rigging.
• Eliminate visual artifacts such as jumping or popping during FK-IK switching in animation pipelines.
• Enable automatic, real-time alignment of FK and IK systems without manual keyframing or interpolation.
• Provide a mathematically robust solution using the Jacobian matrix for joint rotation computation in mixed kinematic systems.

Proposed method

• Formulate the kinematic chain as a system of joint rotations using the Jacobian matrix to relate end-effector motion to joint angles.
• Compute the pseudo-inverse of the Jacobian matrix to solve for joint rotations that align FK and IK configurations.
• Use the pseudo-inverse Jacobian to determine the optimal rotation values for each joint in the chain during transition.
• Apply the computed rotations to the FK system to match the IK target position, ensuring smooth motion continuity.
• Integrate the method into a 3D animation pipeline to enable automatic, real-time FK-IK blending.
• Validate the method by comparing motion transitions with and without the Jacobian-based correction, observing artifact reduction.

Experimental results

Research questions

• RQ1 How can FK and IK systems be seamlessly matched without introducing visual artifacts like popping or jumping?
• RQ2 What mathematical formulation enables accurate and automatic joint rotation computation for FK-IK alignment?
• RQ3 Can the pseudo-inverse of the Jacobian matrix effectively resolve the inverse kinematics problem for smooth FK-IK transitions?
• RQ4 How does the proposed method reduce manual intervention in rigging and animation workflows?
• RQ5 What is the impact of using the Jacobian-based approach on motion continuity and system stability?

Key findings

• The use of the pseudo-inverse Jacobian enables accurate computation of joint rotations required to align FK and IK configurations.
• The method successfully eliminates popping or jumping effects during FK-IK transitions by ensuring smooth, continuous motion.
• The approach allows automatic alignment of FK and IK systems without requiring manual keyframing or interpolation.
• The solution is computationally efficient and suitable for real-time application in 3D animation software.
• The method demonstrates robustness across different joint chain configurations, including arms and other limb-like structures.
• The results confirm that Jacobian-based matching provides a stable and reliable alternative to traditional FK-IK blending techniques.