When   Monday, October 12, 2009   Time   4:00 PM - 5:00 PM  
Where   Technological Instit M 416 2145 Sheridan Rd.   map it
Audience   - Faculty/Staff - Student - Public
Contact   Molly E Scanlon   +1 847 491 5586  
Group   McCormick-Colloquia Engineering Sciences and Applied Mathematics

Applied Math Colloquium

Title: Efficient Hybrid Optimization for Skid-Steering Estimation and Control

Speaker: Professor Todd Murphey, Northwestern University

Abstract:

By design, skid-steered vehicles must violate nonholonomic constraints in order to maneuver. This sliding of the wheel against the ground causes the vehicle to behave discontinuously during a maneuver as well as making the vehicle’s state difficult to estimate; both of these affect a controller’s performance. Estimation of contact state is most naturally framed in the context of switched systems, where the vehicle’s ground interaction is modeled by partitioning the system dynamics into four distinct modes. Thus, as the vehicle maneuvers, the system evolves over some mode sequence, transitioning between modes over some set of switching times. We have developed ?rst- and second-order optimization algorithms for estimating these switching times. Second-order algorithms are in particular of importance because they exhibit quadratic convergence. Moreover, ?rst-order methods often fail to converge on time scales compatible with real-time operation even for relatively simple examples. If noise is introduced into the system, first-order methods can fail entirely, while second-order methods remain efficient. Lastly, estimation of the mode sequence can be achieved by recasting the combinatoric optimization problem as an infinite-dimensional optimization that is easily solved.