Welcome to my Web Page!

Work Information

Ph.D. student in Electrical and Computer Engineering, Carnegie Mellon University

My advisor is Dr. Pradeep K. Khosla. I also work with Dr. Chris Paredis, who has become my second advisor here at Carnegie Mellon University

Key responsibilities

I work in The Millibot Project. This is a research effort being conducted at Carnegie Mellon University under the direction of Pradeep Khosla. The Millibots are one part of a bigger research effort called the CyberScout project. Bob Grabowski and I started the Millibot project during the Spring of 97'.

The CMU Millibots are small, semi-autonomous robots on the 5 - 10 cm scale. They provide support for larger robots in the form of distributed sensors and mobile platforms with small payloads such as cameras, temperature sensors or chemical sensors. Due to the limitation in size, power and cost, the CMU Millibots have limited on-board processing power and are controlled (tasked) by a central host processor. Our research deals with these limitations and tries to overcome them by applying collaboration techniques among a group of Millibots.

The Millibots

My main contribution to this project so far has been the development of a Collaborative Localization System for the CMU Millibot fleet. This system combines aspects of GPS, landmark-based localization, and dead reckoning. The method uses synchronized ultrasound pulses to measure the distances between all the robots on a team and then determines the relative positions of the robots through trilateration. For more information please check out my list of publications.

I also designed and developed a Beacon-Sonar Module for the Millibots. This bifunctional module is a logical extension of the first design described in the papers below. We were already sending and receiving ultrasonic signals for localization purposes, so we thought it would be a waste of acoustic energy if we did not take advantage of the echoes resulting from the localization signals. These echoes provide useful ranging information that the Millibots use for mapping the environment in which they're working, as well as for obstacle detection.

The preliminary version of this circuit was successfully tested recently. You can see a picture of the first Beacon-Sonar Module here. Despite the fact that this module uses the very same circuitry designed for the original localization module, plus 8 additional emitters for sonar purposes, the concept was proved and encouraged further work along these lines.

I just completed the design of the second generation of the Beacon-Sonar Module. It uses a smaller yet better, more sensitive ultrasonic detector as the basic cell for an array of 8 detectors. This allows the new module to measure the time of flights of 8 echoes simultaneously. Besides, the detector array allows the Millibot to determine the relative direction of each incoming pulse. With one single ultrasonic pulse emitted, the detector array is capable of measuring distances to obstacles in eight different directions at the same time. The module consumes less than 25mW of power, which makes it perfectly suited to the Millibots. Here is a picture of the Beacon-Sonar Module installed on top of a Millibot: