CS 326A: Motion Planning ai.stanford.edu/~latombe/cs326/2007/index.htm Criticality-Based Motion Planning (2)
Topics Target finding Information (or belief) state/space Part orientation Sensorless reduction of uncertainty Assembly planning Path space Stereotaxic radiosurgery
Assembly Planning Example
Levels of Problems Parts are assumed free-flying Assembly sequence planning Tools/fixtures are taken into account Entire manipulation system is taken into account Manipulation planning
Assembly Sequence Planning Example of a multi-robot coordination problem, but … 1. Very constrained goal state, but unconstrained initial state Disassembly planning 2. Many dofs, but simple paths Motion space
Various “Interesting” Cases Multi-hand: An assembly on n parts may require up to n hands for its (dis-)assembly [Natarajan] Non-monotonic 2-handed assembly: No single part can be added or removed:
With translations only monotone two-handed With translations only non-monotone, 2-handed monotone, 3-handed With general motions monotone, 2-handed Various “Interesting” Cases
Complexity of Partitioning Assembly partitioning problem: - Given a set of non-overlapping polygons, - Decide if a proper subset of them can be removed as a rigid body without colliding with the other polygons. This problem is NP-complete
OR Gate for u i u j u k
Planning Approaches Generate-and-test: Hypothesize a subassembly and test if it can separated from the rest using contact analysis … But … exponential number of subassemblies: O(2 n ) subassemblies, but only two pairs can be separated
Planning Approaches Generate-and-test Generate-and-test plus caching Non-directional blocking graph (limited to single-step motions) Interference diagram