Lab 12 / 15 · Interactive · real-time
Friction Cone Slicing Bench
Drag the direction angle and wedge count: see how much tangential capability gets sliced off when the cone becomes a polytope.
Cone parameters
What to look at
Dashed circle = true cone, solid = linearized admissible region. The dip is deepest when the pointer sits midway between the two rays.
Inscribed (conservative) vs circumscribed (aggressive) is an engineering trade-off: inscribed underestimates capability, results never slip; circumscribed overestimates and may hand you a grasp that slips. Default: better inside than outside.
What problem this lab solves
Every grasp optimizer must answer: what does approximating the friction cone with an m-wedge polytope cost? The answer's shape is cos(π/m): 4 wedges lose 29%, 8 wedges lose 7.6% — the error vanishes as 1/m² (T12, Theorem 2.3). This plot lets you see what «conservatism» looks like.
How to read the figure
A polar plot: the angle axis is the tangential force direction (relative to the normal), the radius is the allowed tangential force. The dashed circle is the true friction cone (radius μ·f_n, the same everywhere); the solid polygon is the allowed range after m-wedge linearization — it touches the circle only along the rays, and dips deepest between two rays. The pointer is the current direction.
Hands-on
Move the wedge count from 4 to 8 and watch the dip at the worst direction (midway between two rays) recover. Then drag the direction angle to exactly between two rays and read the «underestimate at current direction» — that is precisely where an LP's conservative grasp-optimization results come from.
Done playing — where next
Related topics
Next labs (in play-learn order)
Grasp ε Evaluation Bench
Move fingers, change friction, and watch grasp quality ε shift — and when it hits zero.
7 min
Hertz Contact Calibration Bench
Drag indentation depth and material: see why force goes as depth to the 1.5 power and stiffness grows as you press harder.
6 min
Rolling Holonomy Lab
The ball rolls back to its start, the orientation doesn't — roll one loop and see how far it drifts, and what sets the drift.
6 min