ENGR 330 Lecture

Adaptive Communication Networks for Heterogeneous Teams of Robots
Monday, March 14, 2016 -
3:30pm to 4:30pm
Packard 101
Stephanie Gil (Massachusetts Institute of Technology)
Abstract / Description: 

Multi-robot systems are poised for impact in the future of robotics. Robots will work together on factory floors to make manufacturing more efficient and cost effective, they will coordinate as our teammates during rescue missions in disaster zones, and they will be our eyes in places that we cannot reach such as in space and on other planets. However, in order for multi-robot systems to reach the goals that we've set for them, they must be able to perform one particular task very reliably: information exchange. Information exchange is essential for coordination with humans and with other robots. Currently, wireless communication is largely unreliable in the field; and this unreliability is keeping multi-robot teams from reaching their full impact in the real world. In this talk I will present a novel algorithm for sensing communication with higher accuracy and resolution than what was previously attainable for small agile robot platforms; using only a single off-the-shelf Wi-Fi antenna and local robot motion. This results in the ability of using communication as a high-fidelity sensor. Using this capability, I will derive new control algorithms that can autonomously establish and repair communication links to other robots in the network; even in unknown and unexplored environments. I will show that the ability to use communication, not only to transmit messages but also as a physical signature for each transmitting agent, opens the door to many extensions and applications for multi-robot systems. In addition to establishing adaptive ad-hoc networks, I will discuss an application to cybersecurity in multi-robot teams where each transmitted message by a given robot can be used to provably discern whether that robot is malicious or benign in the context of a Sybil Attack; where malicious robots can spoof or spawn false identities to gain a disproportionate influence in the network. This talk will present several experimental results in hardware for heterogeneous multi-robot systems consisting of iRobot Create and AscTec Hummingbird platforms, in indoor environments. I will conclude by presenting interesting avenues for future research at the intersection of communication and robotics including Wi-Fi enabled accurate indoor positioning systems, communication as a sensing medium for autonomous driving, and new communication channels for natural human-robot interaction.


Stephanie Gil is currently a Research Scientist in the Distributed Robotics Lab at MIT. Her research interests are in multi-robot control, distributed optimization of ad-hoc communication networks, and developing new methods of sensing with Wi-Fi that enable reliable communication for robots in the field, cybersecurity for multi-robot teams, and accurate indoor positioning. Stephanie is the recipient of the National Science Foundation Graduate Research Fellowship, Bell Labs Graduate Research Fellowship, MIT Graduate Research Fellowship and a Deshpande Grant for recent efforts in developing a startup (Ubiety) on accurate indoor positioning. Her past research experience involves work on several NASA projects including being part of the Mars Exploration Rover Team, and participating in an international collaboration for the Future Urban Mobility Project at the National University of Singapore. Stephanie Gil received her B.S. from the Mechanical and Aerospace Engineering department at Cornell University in 2006, her M.S. (2008 under Prof. Brian Williams) and Ph.D (2014 under Prof. Daniela Rus) in the Computer Science and Artificial Intelligence Lab (CSAIL) at MIT.