Ian M. Mitchell

Professor
Computer Science Department
University of British Columbia.
Personal Research & Teaching Contacts:
email: ian.mitchell@ubc.ca
office: ICICS/CS 223
tel: (604) 822-2317 (but I'm not checking it)
fax: (604) 822-5485
mail: 2366 Main Mall, Vancouver, BC, CANADA, V6T 1Z4
Other Academic Pages:
Google Scholar

Courses

Courses numbered 5xx are graduate level. Courses numbered 4xx and below are undergraduate. A note to UBC students: as of the last time I checked, you could count up to two courses from the other category toward your degree (eg grad courses for undergraduates, ugrad courses for grad students). Check with your advisor and the instructor for permission. I am generally quite happy to have students take courses at either level.

I try to keep the graduate course links live, but the undergraduate course links for older courses may point to more recent offerings.


Courses

Courses numbered 5xx are graduate level. Courses numbered 4xx and below are undergraduate. A note to UBC students: as of the last time I checked, you could count up to two courses from the other category toward your degree (eg grad courses for undergraduates, ugrad courses for grad students). Check with your advisor and the instructor for permission. I am generally quite happy to have students take courses at either level.

I try to keep the graduate course links live, but the undergraduate course links for older courses may point to more recent offerings.


Reproducible Research


Research

My research agenda includes:

  • Algorithms for designing and analyzing hybrid and embedded systems.
  • Level set algorithms for time-dependent Hamilton-Jacobi partial differential equations (PDEs).
  • Single pass (FMM, OUM) and sweeping algorithms for static Hamilton-Jacobi PDEs.
  • High quality, publically released implementations of these algorithms.
  • Application of these algorithms to path planning for autonomous and semi-autonomous systems.

For those interested in experimenting with Level Set Methods for solving time-dependent Hamilton-Jacobi PDEs, such as those that arise in dynamic implicit surfaces and reachability, I have released:

A Toolbox of Level Set Methods

I do not yet have any public code for direct approximation of static (time-independent) Hamilton-Jacobi PDEs, such as those that arise in robotic path planning. However, here are some slides discussing the relationship between path planning and Hamilton-Jacobi PDEs, various classes of algorithms for approximating the resulting PDEs, and a few extensions on which I have worked: Dynamic Programming Algorithms for Planning and Robotics in Continuous Domains and the Hamilton-Jacobi Equation. These slides were presented at the Fast Marching Method: Theoretical Underpinnings and Applications to Robotics tutorial at IROS 2008.


Selected publications:

Many of my academic publications are available online. Where permitted by the publishers, I have provided links to self-archived post-prints (if I have one) as well as the published versions (the latter usually require subscriptions). If the paper you want is not hyperlinked or if you are unable to access the publisher's web site, please email me and I may be able to provide a copy by other means.

2017

2016

2015

2014

2013

2012

2011

2010

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998


Books

Technical Reports and Other Publications


Talks

Here are some talks I have given which do not directly correspond to research publications.

Links

Local Research Groups:

Conferences and Workshops:

Useful Stuff:


Postdoctoral Work

Selected Topics
Constrained Path Planning in continuous state spaces (images at right).
Softwalls (preventing the use of commercial aircraft as weapons).
Performed at (September 2002 to July 2003)
80% in the Robotics Lab, EECS Department, University of California at Berkeley
20% in the Physically Based Modeling Group, Computer Science Department, Stanford University.
Postdoctoral Supervisors:
Shankar Sastry from the EECS Department
Ronald Fedkiw from the Computer Science Department

avi mpg

Doctoral Work

Selected Topics
Calculating reachable sets for Continuous and Hybrid Systems
Level Set Methods and Algorithms
Performed in (September 1997 to August 2002)
SCCM program, Stanford University
Doctoral Supervisors:
Claire Tomlin from the Aero/Astro Department
Ronald Fedkiw from the Computer Science Department

Growing Set: avi mpg
Final Set: avi mpg

Created 19 August 2003.
Last updated 07 June 2019.
Created and maintained by Ian M. Mitchell.