This page is no longer updated.
Please visit my new page instead.

Jonathan Gammell

Ph.D. (Autonomous Space Robotics Lab, University of Toronto)
M.A.Sc. (Space Robotics Group, University of Toronto)
B.A.Sc. (Mechanical Engineering, University of Waterloo)

Institute for Aerospace Studies
University of Toronto
4925 Dufferin Street, Room 150B
Toronto, Ontario, Canada
M3H 5T6
jon.gammell [at] utoronto.ca
office: +1 416-667-7730
fax: +1 416-667-7799
skype: jdgammell
Google Scholar
arXiv

Jonathan Gammell

Research • Education • Teaching • Field Deployments • Collaborations • Publications

Research

Robotics is important because it can be applied to a wide-range of disciplines. While my focus has been on autonomous mobile robots in unstructured environments, I have been fortunate enough to also work on projects in a variety of other disciplines. Recently, my robotics research has focused on developing informed techniques to improve optimal sampling-based planning. This has led to some new ideas that I find exciting, specifically that solving the optimal planning problem generally requires direct-sampling of an informed subset and that sampling-based planners can be viewed as graph-searches of implicit random geometric graphs. More information about these ideas are available in the Informed RRT*, and Batch Informed Trees (BIT*) work, respectively.

Informed RRT*

Informed RRT* is an improvement to the RRT* algorithm that increases the rate at which the found solution converges to the optimum. Developed specifically for problems seeking to minimize path-length, this work uses a method to directly sample the subset of a planning problem that contains all possible improvements to a given solution. In doing so, it continues to consider all relevant paths without requiring any additional parameters. This direct sampling is actually required to generally solve the planning problem, as the probability of improving a given solution goes to zero as the size of the problem domain, or the number of dimensions, increases. For a 6DOF arm, even an intelligent sample-rejection technique will limit the probability of improvement at each iteration to ~8%, while direct sampling imposes no such limit.

This work was presented at the 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) in Chicago. The paper is available from here or arXiv (arXiv:1404.2334 [cs.RO]), along with 1, 2 short associated proofs (arXiv: arXiv:1403.7664 [math.ST] and arXiv:1404.1347 [math.ST]). Short videos of the algorithm and results are also available on YouTube. An OMPL version is available from the official repository and will be included in upcoming binary releases.
Coauthors: Siddhartha Srinivasa, Tim Barfoot.

Batch Informed Trees (BIT*)

Batch Informed Trees (BIT*) is an anytime optimal sampling-based planner that uses the principles of dynamic programming to search the implicit random geometric graph (RGG) given by batches of samples drawn from the problem domain. To do this efficiently, it uses a heuristic much like A* to prioritize both expansion towards the goal and high-quality paths. Initial experimental results show that it outperforms existing optimal sampling-based planning algorithms (e.g., RRT*, FMT*) in terms of computational cost to find equivalent results while still providing an anytime solution.

This work was presented at the 2015 IEEE Conference on Robotics and Automation (ICRA) in Seattle, May 26-31. The paper is available from here or arXiv (1405.5848 [cs.RO]) and a short video of results is available on YouTube. An OMPL version is available from the official repository and will be included in upcoming binary releases.
Coauthors: Siddhartha Srinivasa, Tim Barfoot.

Education

Teaching

  • ESC203: Engineering, Society and Critical Thinking (University of Toronto) - Teaching Assistant (Fall 2009 - 2014)

  • Sargasso Cloud Workshop (The Royal Danish Academy of Fine Arts. Copenhagen, Denmark) - Teaching Assistant (August 2009)

  • Epithelium Design Studio Workshop (Pratt Institute. Brooklyn, New York) - Teaching Assistant (September 2008)

Field Deployments

One of the most exciting parts of my research here at ASRL is having the opportunity to test the real-world performance of rover systems and algorithms in real mission scenarios on field deployments and analogue missions. The tests range in length from just a week to the better part of a month and can consist of as little as a single team of 4 people or multiple teams totalling almost fifty people. Included in each deployment description below is the rover platform used in the mission, the general location of the field test and a link to the specific location on Google Maps. If you would like to see the complete deployment map in a separate window, you can go directly to Google Maps.

Collaborations

In addition to my own research, I enjoy working on projects with friends or colleagues in their areas of interest. It's amazing what people in other fields can create when provided with some simple engineering support and it's rewarding to be able to contribute in any small way.

  • Flicker - Eric Bury + Farid Noufaily (DesCours 2010, New Orleans, Louisiana)

  • Sargasso Cloud - Philip Beesley Architect Inc. (The Royal Danish Academy of Fine Arts, Copenhagen, Denmark)

  • Epithelium - Philip Beesley Architect Inc. (Pratt Institute of Art and Design, Brooklyn, New York)

Flicker - Eric Bury + Farid Noufaily

Publications

Due to copyright issues, several of the PDFs below must be password protected. Most can also be found on Google Scholar, and as preprints on arXiv. If you would like access to a paper you cannot find elsewhere, please do not hesitate to email me.

Journal and Magazine Articles

  1. Moores J, Francis R, Mader M, Osinski G R, Barfoot T D, Barry N, Basic G, Battler M, Beauchamp M, Blain S, Bondy M, Capitan R D, Chanou A, Clayton J, Cloutis E, Daly M, Dickinson C, Dong H, Flemming R, Furgale P, Gammell J, Gharfoor N, Hussein M, Grieve R, Henrys H, Jaziobedski P, Lambert A, Leung K, Marion C, McCullough E, McManus C, Neish C D, Ng H H, Ozaruk A, Pickersgill A, Preston L J, Redman D, Sapers H, Shankar B, Singleton A, Souders K, Stenning B, Stooke P, Sylvester P, and Tornabene L. "A Mission Control Architecture for Lunar Sample Return as Field Tested in an Analogue Deployment to the Sudbury Impact Structure." Advances in Space Research, special issue on Lunar Exploration, 50(12): 1666-1686, 2012. doi:10.1016/j.asr.2012.05.008. (pdf, google)

Conference Papers

Full-Paper-Refereed

  1. Choudhury S, Gammell J D, Barfoot T D, Srinivasa S S, Scherer S. "Regionally Accelerated Batch Informed Trees (RABIT*): A Framework to Integrate Local Information into Optimal Path Planning." In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pp. 4207-4214. Stockholm, Sweden, 16-21 May 2016. doi:10.1109/ICRA.2016.7487615. (pdf, google)

  2. Gammell J D, Srinivasa S S, Barfoot T D. "Batch Informed Trees (BIT*): Sampling-based Optimal Planning via the Heuristically Guided Search of Implicit Random Geometric Graphs." In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pp. 3067-3074. Seattle, WA, USA, 26-31 May 2015. doi:10.1109/ICRA.2015.7139620. (video, pdf, arXiv:1405.5848 [cs.RO], google)

  3. Gammell J D, Srinivasa S S, and Barfoot T D. "Informed RRT*: Optimal Sampling-based Path Planning Focused via Direct Sampling of an Admissible Ellipsoidal Heuristic." In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 2997-3004. Chicago, IL, USA, 14-18 Sept. 2014. doi:10.1109/IROS.2014.6942976. (video 1, video 2, pdf, arXiv:1404.2334 [cs.RO], google)

  4. Gammell J D, Tong C, and Barfoot T D. "Blinded by the Light: Exploiting the Deficiencies of a Laser Rangefinder for Rover Attitude Estimation." In Proceedings of the 10th Conference on Computer and Robot Vision (CRV), pp. 144-150. Regina, SK, Canada, 28-31 May 2013. doi:10.1109/CRV.2013.37. (pdf, google)

  5. Berczi L-P, Gammell J D, Tong C, Daly M, and Barfoot T D. "A Proof-of-Concept, Rover-Based System for Autonomously Locating Methane Gas Sources on Mars." In Proceedings of the 10th Conference on Computer and Robot Vision (CRV), pp. 29-36. Regina, SK, Canada, 28-31 May 2013. doi:10.1109/CRV.2013.16. (pdf, google)

  6. Gammell J D, Tong C H, Berczi P, Anderson S, Barfoot T D, and Enright J. "Rover Odometry Aided by a Star Tracker." In Proceedings of the IEEE Aerospace Conference, pp. 1-10. Big Sky, MT, USA, 2-9 March 2013. doi:10.1109/AERO.2013.6496953. (pdf, google)

Abstract-Refereed

  1. Gammell J D, Srinivasa S S, and Barfoot T D. "BIT*: Batch Informed Trees." In Proceedings of the Information-based Grasp and Manipulation Planning Workshop, Robotics: Science and Systems (RSS). Berkeley, CA, USA, 13 July 2014. (pdf, poster)

  2. Barfoot T D, McManus C, Anderson S, Dong H, Beerepoot E, Tong C H, Furgale P, Gammell J D, and Enright J. "Into Darkness: Visual Navigation based on a Lidar-Intensity-Image Pipeline." In Proceedings of the International Symposium on Robotics Research (ISRR). Singapore, 16-19 Dec 2013. (pdf)

  3. Barfoot T D, Stenning B E, Gammell J D, Tong C H, McManus C, Berczi L P, Osinski G R, Daly M, and Dickinson C. "Place Revisiting for Planetary Rovers: An Enabling Technology and Field Testing of Three Mission Concepts." In Proceedings of the Planetary Rovers Workshop, IEEE International Conference on Robotics and Automation (ICRA). Karlsruhe, Germany, 10 May 2013. (pdf)

  4. Merali R S, Tong C, Gammell J, Bakambu J, Dupuis E, and Barfoot T D. "3D Surface Mapping Using a Semi-Autonomous Rover: A Planetary Analog Field Experiment." In Proceedings of the 2012 International Symposium on Artificial Intelligence, Robotics and Automation in Space (i-SAIRAS). Turin, Italy, 4-6 September 2012. (pdf)

  5. Stenning B, Osinski G R, Barfoot T D, Basic G, Beauchamp M, Daly M, Francis R, Furgale P, Gammell J, Ghafoor N, Jasiobedzki P, Lambert A, Leung K, Mader M, Marion C, McCullough E, McManus C, Moores J, and Preston L. "Planetary Surface Exploration Using a Network of Reusable Paths." In Proceedings of the 2012 International Symposium on Artificial Intelligence, Robotics and Automation in Space (i-SAIRAS). Turin, Italy, 4-6 September 2012. (pdf)

  6. Marion C L, Osinski G R, Abou-Aly S, Antonenko I, Barfoot T, Barry N, Bassi A, Battler M, Beauchamp M, Bondy M, Blain S, Capitan R, Cloutis E, Cupelli L, Chanou A, Clayton J, Daly M, Dong H, Ferriere L, Flemming R, Flynn L, Francis R, Furgale P, Gammell J, Garbino A, Ghafoor N, Grieve R A F, Hodges K, Hussein M, Jasiobedzki P, Jolliff B L, Kerrigan M C, Lambert A, Leung K, Mader M M, McCullough E, McManus C, Moores J, Ng H K, Otto C, Ozaruk A, Pickersgill A E, Pontefract A, Preston L J, Redman D, Sapers H, Shankar B, Shaver C, Singleton A, Souders K, Stenning B, Stooke P, Sylvester P, Tripp J, Tornabene L L, Unrau T, Veillette D, Young K, and Zanetti M. "A Series of Robotic and Human Analogue Missions in Support of Lunar Sample Return." In Proceedings of the 43rd Lunar and Planetary Science Conference (LPSC). The Woodlands, TX, USA, 19-23 March 2012. (pdf)

  7. Pickersgill A E, Osinski G R, Beauchamp M, Marion C, Mader M M, Francis R, McCullough E, Shanker B, Barfoot T D, Bondy M, Chanou A, Daly M, Dong H, Furgale P, Gammell J, N N G, Hussein M, Jasiobedski P, Lambert A, Leung K, McManus C, Ng H K, Pontefract A, Stenning B, Tornabene L L, Tripp J, and Team I. "Scientific Instrumentation for a Lunar Sample Return Analogue Mission." In Proceedings of the 43rd Lunar and Planetary Science Conference (LPSC). The Woodlands, TX, USA, 19-23 March 2012. (pdf)

  8. Stenning B, Osinski G R, Barfoot T D, Basic G, Beauchamp M, Daly M, Dong H, Francis R, Furgale P, Gammell J, Ghafoor N, Lambert A, Leung K, Mader M, Marion C, McCullough E, McManus C, Moores J, and Preston L. "Planetary Surface Exploration Using a Network of Reusable Paths." In Proceedings of the 43rd Lunar and Planetary Science Conference (LPSC). The Woodlands, TX, USA, 19-23 March 2012. (pdf)

  9. Moores J, Francis R, Barfoot T, Barry N, Basic G, Battler M, Beauchamp M, Blain S, Bondy M, Capitan R D, Chanou A, Clayton J, Cloutis E, Daly M, Dickinson C, Dong H, Flemming R, Furgale P, Gammell J, Ghafoor N, Hussein M, Grieve R, Henrys H, Jaziobedski P, Lambert A, Leung K, Mader M, Marion C, McCullough E, McManus C, Neish C, Ng H, Ozaruk A, Pickersgill A, Preston L, Redman D, Sapers H, Shankar B, Singleton A, Souders K, Stenning B, Stooke P, Sylvester P, Tornabene L, and Osinski G. "Mission Operations Design for Lunar Sample Return as Field-Tested in an Analogue Deployment to the Sudbury Impact Structure." In Proceedings of the European Planetary Science Congress (EPSC-DPS). Nantes, France, October 2011. p. 1728. (pdf)

  10. Saini R, Jandric Z, Gammell J, Baur C, Mentink S A and Tuggle D. "Manufacturable MEMS miniSEMS." In International Conference on Micro- and Nano-Engineering, September 2005. p. 734-735.

  11. Jandric Z, Saini R, Gammell J, Tuggle D and Skidmore G. "Scanning Electron Microscopy - Next Generation." In Proceedings of 10th International Commercialization of Micro-Nano Systems (COMS) Conference, August 2005. pp. 493-499

Technical Reports

  1. Gammell J D, Barfoot T D. "The Probability Density Function of a Transformation-based Hyperellipsoid Sampling Technique". Technical Report TR-2014-JDG004, University of Toronto, 2014. arXiv:1404.1347 [math.ST]. (pdf)

  2. Gammell J D, Srinivasa S S, Barfoot T D. "On Recursive Random Prolate Hyperspheroids". Technical Report TR-2014-JDG002, University of Toronto, 2014. arXiv:1403.7664 [math.ST]. (pdf)

Thesis

  1. Gammell J D. "Informed Anytime Search for Continuous Planning Problems". Ph.D. Thesis, University of Toronto, 2017. (pdf)

  2. Gammell J D. "Evaluation of a Mobile Platform for Proof-of-concept Autonomous Site Selection and Preparation". M.A.Sc. Thesis, University of Toronto, 2010. doi:1807/25591. (pdf)

Scholarly Presentations

  1. Gammell J D. "Informed Anytime Search for Continuous Planning Problems," Departmental Doctoral Seminar. Institute for Aerospace Studies, University of Toronto, Toronto, ON, Canada. 31 January 2017.

  2. Gammell J D. "Informed Sampling-based Planning," Invited Talk. Personal Robotics Lab, Carnegie Mellon University (CMU), Pittsburgh, PA, USA. 4 September 2014.

  3. Gammell J D. "Using ROS on Field Robotic Experiments in Remote Locations," Ninth Conference on Computer and Robot Vision (CRV 2012) ROS Workshop, Toronto, ON, Canada. 27 May 2012.

  4. Gammell J D, and Tong C. "Using ROS on Field Robotic Experiments in Remote Locations," ROSCon 2012 Developer Conference, St. Paul, MN, USA. 19 May 2012.

  5. Gammell J D. "Development of a Mobile Platform for Proof-of-Concept Autonomous Site Selection and Preparation," Invited Talk. Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland. 22 November 2010.

  6. Gammell J D. "Evaluation of a Mobile Platform for Proof-of-Concept Autonomous Site Selection and Preparation," Deparmental Master's Seminar. Institute for Aerospace Studies, University of Toronto, Toronto, ON, Canada. 21 September, 2010.