Title: Fast Implementation of Lemke
1Fast Implementation of Lemkes Algorithm for
Rigid Body Contact Simulation
John E. Lloyd
Computer Science Department University of British
Columbia Vancouver, Canada
2Applications mechanical simulation, animation,
haptics
Haptics requires speed (1 Khz) and accuracy
3Extended contact can result in many contact points
4Contributions
Most exact solution method is Lemkes
algorithm with expected complexity
- Use problem structure to speed up solution
- Reduce complexity to nearly
- complexity for fixed number of bodies
number of contacts, number of
bodies
5Problem formulation
6Constraints
7Results in a Linear Complementarity Problem (LCP)
8Solving Contact LCPs
- Iterative techniques includes impulse methods
Mirtich Canny 95, Guendelman 03
gt Accuracy, convergence?
- Pivoting methods Lemkes algorithm Anitescu
Potra 97, Stewart Trinkle 96 gt
Speed, robustness?
9Pivoting exchange subsets of z and w
Generally, one variable exchange per pivot
10Once per pivot compute
- Involves solving
- Complexity , and typically
pivots - Hence total expected complexity
11Peg in hole test case
12How to improve performance?
- 1 Solve more efficiently
- 2 Reduce the number of pivots
131 Solving
Ignoring Lemke covering vector in this discussion
14This reduces the system to
15This yields the final system
- Reduced matrix has size
- Hence per-pivot computation is
- So total expected complexity
162 Reducing the number of pivots
17So start with a frictionless LCP
18Ideally, final system rank is
19Results peg in hole
Standard Structural Reduced
20Results sample contacts
Standard Structural Reduced
21Results block stack
Standard Structural Reduced
22Results
23Conclusions Improved pivoting method for contact
simulation
- Fast exploit problem structure
- Better complexity nearly
- for fixed number of bodies
- Efficient no need to compute
- More robust smaller system to solve each pivot
http//www.cs.ubc.ca/lloyd/fastContact.html
24(No Transcript)
25Future Work
- Exploit temporal coherence (give solver an
advanced starting point) - More efficient solution for reduced equation
- Robust pivot selection (minimum ratio test)
26LCP matrix can be quite large
27Larger number of needed for accurate
friction computation
f
v
28Closeup sample contacts