Events Calendar

AME Seminar

Speaker: Nima Fazeli, MIT

Talk Title: Towards Robotic Manipulation, Understanding the World Through Contact

Abstract: Why is robotic manipulation so hard? As humans, we are unrivaled in our ability to dexterously manipulate objects and exhibit complex skills seemingly effortlessly. Recent research in cognitive science suggests that this ability is driven by our internal representations of the physical world, built over a life-time of experience. Our predictive ability is complemented by our senses of sight and touch, intuitive state-estimation, and tactile dexterity. Given the complexity of human reasoning, skill, and hardware, it is not surprising that we have yet to replicate our abilities in robots. In order to bridge this gap, we must develop robotic systems that build their understanding and interpretation of the physical world through contact. Using experiments as tools, these Galilean Robots will distill their experiences into models of the physical world.

In this talk, I will present some of my work spanning the spectrum of analytical to fully data driven methodologies for model building and inference through contact. I believe that Galilean Robots need to master tools from this spectrum for intelligent and dexterous manipulation. First, I will discuss a methodology for the inference of contact forces and system parameters of rigid bodies systems making and breaking contact. I will then touch on data augmented contact models for controls as a medium between analytical and data driven techniques. I will show how a robot can learn the physics of playing Jenga using a hierarchical learning methodology purely from data. I will conclude the talk by providing perspectives on building Galilean Robotic systems that embody intelligent manipulation.

Nima Fazeli is a PhD student with the Mechanical Engineering Department at MIT, working with Prof. Alberto Rodriguez. His research focuses on enabling intelligent and dexterous robotic manipulation by developing novel tools combining analytical methods, machine learning, and cognition/AI. During his PhD, Nima has developed inference algorithms for robotic systems undergoing frictional contact, performed empirical evaluations of contact models, demonstrated data-augmented contact models for manipulation, and developed a robotic system capable of learning the physics of playing Jenga using a hierarchical learning methodology. Nima received his masters from the University of Maryland at College Park where he spent most of his time developing analytical and data-driven models of the human (and, on occasion, swine) arterial tree together with novel inference algorithms to diagnoses cardiovascular diseases. His research has been supported by the Rohsenow Fellowship and featured in outlets such as CBS, CNN, and the BBC. He looks forward to robots playing and learning alongside his grandchildren.

Thursday, April 4, 2019
11:00 AM
The Laufer Library (RRB 208)
Refreshments will be served at 10:45 AM.

Host: AME Department

Location: Robert Glen Rapp Engineering Research Building (RRB) - 208 (Laufer Library)

Audiences: Everyone Is Invited

Contact: Tessa Yao