Manipulation in Unstructured Settings

When:
November 10, 2015 @ 6:00 pm – 9:00 pm America/New York Timezone
2015-11-10T18:00:00-05:00
2015-11-10T21:00:00-05:00
Where:
iRobot
8 Crosby Dr
Bedford, MA 01730
USA

Robotics and Automation Society

Manipulation in Unstructured Settings – Presented by Joseph Romano, Ph.D.

Nominations for Chapter Officer Positions will also be accepted at this meeting! Contact chair@robotics-boston.org if you have questions about volunteering for the Chapter, or the upcoming election!

Doors open: 6:00 P.M.
Presentation: 6:30 P.M.
Dinner: 8:00 P.M.

Industrial tasks such as welding, painting, and part transport were some of the first commercially successful uses of robots. A significant contributor to that success has been the ability to tightly control the variability in these environments with things like limited object sets, fixtures/jigs, and safety cages. Advances in hardware (low-cost sensors, high-precision low-volume manufacturing) and software (algorithmic research in perception, control and planning) has renewed interest in applying robot manipulation to problems with far less structure. These modern efforts include things like handling parts where the geometric or material properties are not well known, or accomplishing tasks without an a priori detailed model of the world.

The challenges that arise with automation in unstructured settings are often surprising and unexpected. For example, motion speed and accuracy of robotic manipulation systems outpaced human capabilities long ago, but robots still usually lag far behind on task-level throughput and reliability. The subtleties of human manipulation that allow us to complete everyday tasks are often poorly understood, and are challenging to translate into automated subroutines. Even the seemingly simple task of grasping and transporting an object remains a major research challenge when lacking detailed prior information about the object and environment properties to design around.

This presentation will give examples of robotic systems applied to industrial problems, and discuss interesting challenges that arise when assumptions about the world’s structure and reality no longer align. This will serve as a motivator for discussing various modern techniques making their way from academia to industry for handling unstructured settings.

Joseph Romano, Ph.D.

Joe is a Research Scientist developing next-generation robotic platforms at Berkshire Grey Inc. in Waltham, Massachusetts. Prior to BG, Joe was part of the engineering team that brought Rethink Robotics Baxter Robot to life and part of the advanced research team at Kiva Systems (Amazon Robotics) where he designed the Amazon Picking Challenge academic contest . He received his Masters/PhD from the University of Pennsylvania’s GRASP laboratory focusing on algorithms for robotic tactile manipulation and virtual haptic rendering, and his BS at Johns Hopkins University working in surgical robotics. Joe’s interests span signal-processing, control, and planning strategies that allow robots to accomplish delicate sensor-aware manipulation tasks.

Meeting Location: iRobot, 8 Crosby Drive, Bedford, MA 01730

Unhosted Dinner: Flatbread Company, 213 Burlington Road, Bedford, MA 01730

Have more questions? Want to share a drink with the speaker? Want to network with fellow engineers and professionals? Just want to chat about the current goings-on in Robotics, or technology in general? Join us for dinner, where you can talk about Robotics in a more casual setting!

General Information: This and other RAS meetings are open to the general public. For more information about the RAS Boston Chapter, contact Chapter Chair Ryan Pettigrew at chair@robotics-boston.org or visit http://www.robotics-boston.org/.

To assist us in planning this meeting, please pre-register at http://www.ieeeboston.org/Register/.