Form Closure

Form closure is achieved when an object is completely immobilized by contact with rigid bodies (e.g., gripper fingers or fixtures), preventing any infinitesimal motion. This is a necessary condition for a secure grasp without relying on friction. The analysis treats the object and contacts as perfectly rigid, evaluating mobility based solely on geometry. It is mathematically defined using screw theory and wrench spaces, where the set of contact wrenches must positively span the entire space of possible motions.

A key distinction from force closure is that form closure does not require friction. The grasp is stable even if contact surfaces are perfectly smooth. This makes it a stronger, more robust condition, often sought for critical fixturing in manufacturing or for grasping very slippery objects. The analysis assumes only that contacts can apply forces normal to the object surface (i.e., unilateral constraints). Because it ignores tangential friction forces, a form-closure grasp remains stable under external disturbances that might break a friction-dependent grasp.

The minimum number of frictionless contact points needed for form closure depends on the object's degrees of freedom:

• In 2D (planar grasp): At least 4 points are required to immobilize a 3-DOF object (x, y, rotation).
• In 3D (spatial grasp): At least 7 points are required to immobilize a 6-DOF object. These are necessary but not sufficient conditions. The points must be arranged such that any motion of the object would require it to penetrate at least one of the rigid finger bodies, violating the non-penetration constraint. The placement is highly non-intuitive and is a classic problem in computational geometry.

Determining if a set of contacts yields form closure is solved using linear programming. The core question: Can a set of non-negative contact forces (normal to the surfaces) resist any arbitrary external wrench? This is formulated as a feasibility problem in wrench space. Grasp synthesis algorithms search for these contact configurations, often using sampling or optimization on a 3D model of the object. The Grasp Wrench Space (GWS) is a common analytical tool, where form closure exists if the convex hull of primitive contact wrenches contains the origin in its interior.

Form closure is paramount in mechanical fixturing (e.g., in CNC machining) where a part must be held absolutely stationary against cutting forces. Modular fixturing systems and phase-changing jigs (e.g., using granular jamming) are designed to achieve this. In robotics, it's critical for precision assembly tasks where even microscopic slip is unacceptable, or for handling objects with unknown or low friction coefficients. It is the underlying principle for non-prehensile manipulation like pushing an object against a fence to achieve a predictable, immobilized orientation.

Form closure is often overly restrictive for general-purpose robotic grasping. Achieving the required number and precise placement of frictionless contacts with a standard parallel-jaw or 3-finger gripper is frequently impossible for arbitrary objects. Therefore, most practical robotic grasps rely on force closure, which utilizes friction and requires fewer contacts (e.g., 2 fingers with friction can achieve force closure in 2D). Form closure is a subset of force closure; all form-closure grasps are force-closure grasps, but not vice-versa. This relationship is foundational to grasp quality metrics.

A classic textbook example. A rigid cube is placed between the two parallel, flat faces of a gripper. When the gripper closes, the four contact points (two per face) constrain all six degrees of freedom of the cube:

• Translations along the x, y, and z axes are prevented.
• Rotations about all three axes are prevented. This is a form-closed grasp because friction is not required for immobilization; the geometry of the contacts alone is sufficient. It is often used as a benchmark for grasp analysis algorithms.

Industrial vises, jigs, and fixtures are engineered to achieve form closure for manufacturing. A milling machine vise uses fixed and moving jaws with precisely machined surfaces to hold a workpiece. A 3-2-1 locating principle is often employed:

• 3 points constrain translation and rotation in one plane.
• 2 points constrain in a perpendicular plane.
• 1 point constrains in the final direction. This systematic use of rigid contacts ensures the part cannot move during high-force machining operations, guaranteeing precision and repeatability without relying on clamping friction alone.

Placing a polygonal object (like a hex nut) inside a perfectly matching concave socket achieves form closure. The internal walls of the socket make contact with multiple faces of the object.

• Key Insight: The object's vertices and edges are constrained by the enclosing surfaces.
• Real-world analog: A socket wrench provides form closure on a hexagonal bolt head, allowing high torque transfer. The motion is completely constrained within the socket; the bolt cannot tilt or translate. This contrasts with an adjustable wrench, which relies on friction and is prone to slippage (force closure).

The interlocking pieces of a jigsaw puzzle are a macroscopic, everyday example of mutual form closure. The protrusions (tabs) and cavities (blanks) of adjacent pieces geometrically constrain each other's in-plane motion.

• Constraints: Each piece is prevented from translating in the plane or rotating out of the plane by its neighbors.
• System-level closure: The entire puzzle achieves global form closure when the border pieces are locked in place by the frame or adjacent pieces. This demonstrates how local geometric constraints can propagate to immobilize a complex assembly.

This thought experiment highlights the difference between form closure and force closure. Imagine trying to immobilize a smooth, rigid sphere between two flat, parallel plates.

• Without Friction: The sphere can always rotate and slide. It has form closure? No. The contacts do not prevent rotation about the axis aligned with the line connecting the contact points.
• With Friction: Sufficient normal force can generate enough frictional force to prevent sliding, achieving force closure. This example is critical for grasp planning: many common grasps (like a parallel-jaw gripper on a cylinder) are not form-closed and require friction to be stable.

Achieving form closure with a dexterous, multi-fingered hand is a complex planning problem. The Salisbury Hand (Stanford/JPL Hand) was designed with this principle in mind.

• Four-Finger Grasp on a Sphere: By positioning three fingers in a triangle and the fourth opposite, the hand can achieve form closure on a sphere, constraining all rotational degrees of freedom.
• Planning Challenge: The algorithm must find finger placements such that the contact normals positively span the space of all possible wrenches (forces and torques). This is computationally intensive but guarantees stability independent of friction coefficients, making it highly robust for critical tasks.

Force closure is a stricter, more practical condition for a stable grasp than form closure. It considers the ability of contact forces, including friction, to resist any external wrench (force and torque) applied to the object.

• Key Difference: Form closure is purely geometric and frictionless. Force closure leverages friction, allowing grasps with fewer contact points (e.g., a three-fingered grip on a sphere).
• Mathematical Basis: Achieved when the set of possible contact wrenches positively spans the entire wrench space.
• Real-World Relevance: Almost all practical robotic grasps rely on force closure, as friction is nearly always present and allows for more efficient gripper designs.

Grasp planning is the algorithmic process of determining where and how a robotic gripper should contact an object to achieve a stable and task-functional grasp.

• Inputs: A 3D model or point cloud of the object, the gripper's kinematic model, and task constraints (e.g., required approach direction).
• Outputs: A set of candidate grasp poses (positions and orientations for the gripper).
• Relation to Form Closure: Analytical grasp planners often use form closure and force closure as quality metrics to evaluate and rank candidate grasps. A planner might filter for grasps that are first-order form closed as a baseline for stability.

Compliant grasping is a control or mechanical strategy where a gripper or manipulator can passively or actively yield to contact forces, often to achieve stable grasps on uncertain or fragile objects without relying on perfect geometric closure.

• Passive Compliance: Uses soft materials, springs, or underactuated mechanisms in the gripper itself to conform to object shape.
• Active Compliance: Uses force/torque sensing and control algorithms (like impedance control) to modulate grip force and position in response to sensor feedback.
• Contrast with Form Closure: Form closure assumes rigid bodies and perfect geometric constraints. Compliant grasping is often used when these assumptions break down, supplementing or replacing the need for perfect geometric immobilization.

Caging is a pre-grasping strategy where an object is partially enclosed by a robot's fingers or workspace such that its mobility is topologically limited, even if it is not yet form or force closed.

• Principle: The object cannot escape the "cage" formed by the gripper links without a large, non-infinitesimal motion.
• Two-Phase Strategy: Caging is often used as a reliable initial step: 1) Cage the object with a simple motion, 2) Close the fingers to achieve a final form or force-closed grasp.
• Advantage: More robust to perception and control errors than trying to achieve a precise form-closed grasp in a single motion from a distance.

The friction cone is a critical geometric model used in analyzing force closure. It represents the set of all possible contact force vectors that can be applied at a point without causing slip, given a coefficient of friction.

• Visualization: A cone whose central axis is the surface normal at the contact point. The angle of the cone is arctan(μ), where μ is the coefficient of friction.
• Role in Closure Analysis: For a grasp to be force closed, there must exist contact forces within their respective friction cones whose resultant can balance any external wrench.
• Link to Form Closure: In the special case where μ = 0 (frictionless), the friction cone collapses to a line along the surface normal. This is the assumption underlying pure form closure analysis.

Task and Motion Planning (TAMP) is an integrated approach that combines high-level symbolic task planning (deciding what to do) with low-level geometric motion planning (deciding how to move) to solve complex manipulation problems.

• Hierarchy: A TAMP solver might decide a sequence like "move to bottle, grasp bottle, move to cup, pour." Each symbolic action (e.g., "grasp") requires a feasible geometric motion plan.
• Where Form Closure Fits: The "grasp" action's feasibility is often conditioned on the existence of a stable grasp. The planner might call a grasp planner to find a form- or force-closed configuration as a precondition for the action's success.
• Holistic View: TAMP frameworks treat form closure not as an isolated goal, but as a geometric constraint embedded within a long-horizon, logic-based task plan.