References
References available upon request.
Last updated on Friday, June 16, 1997
University of Maryland
Institute for Advanced Computer
Studies
A.V.
Williams Building
College Park, MD 20742
Phone: (301) 405-6766 Fax: (301) 314-9658
Developed DAMN: A Distributed Architecture for Mobile Navigation. for control of mobile robots systems in complex, uncertain domains with real-time demands. Behaviors communicate with a central command arbiter to collectively determine a robot's actions, voting for and against a range of possible actions according to their objectives and constraints; the arbiter is then responsible for fusing the behaviors' votes to generate action that combines multiple goals and priorities. DAMN has been used to combine various subsystems of differing capabilities on several mobile robots, creating systems for vehicle navigation and active sensor control, thereby contributing significantly to the successes of the Navlab project and the ARPA/OSD Demo II program. Vehicles under the control of DAMN drive at highway speeds, navigate across stretches of off-road terrain some kilometers in length, cooperate with other robotic vehicles, and accept commands from human supervisors. DAMN is used at Carnegie Mellon University, Georgia Institute of Technology, Lockheed Martin, Hughes Research Labs, and the Army Research Labs.
Specialization in Mobile Robots perception, planning, and control. Projects included design of robot for inspection and repair of space shuttle heat tiles, and design and implementation of factory job-shop scheduler. Courses taken include knowledge-based systems, cognitive science, computer vision, operations research, and seminars in intelligent agents and reinforcement learning.
As a teaching assistant for the "Introduction to Mobile Robots" graduate level course, prepared and presented lectures, led discussion groups, designed and directed an extensive class project, and assisted in the preparation and grading of exams. Also participated in the Preparing for a Faculty Career seminar series and the University Teaching Center Seminars for Graduate Students.
Concentration in Artificial Intelligence. Senior thesis topic was extension of Truth Maintenance Systems to include temporal reasoning and multi-valued logics. Graduate level courses taken: expert systems and "Society of Mind"; other courses include artificial intelligence, cognitive psychology, distributed systems, theory of computation, compiler design, computer architectures, formal logic, electronic circuits and systems, and signal theory.
Director of the Autonomous Mobile Robots Laboratory, affiliated with the University's Center for Automation Research and Institute for Systems Research; directing research projects in mobile robot navigation, supervising both graduate and undergraduate students. Extending DAMN architecture to include new capabilities such as task-level and time-dependent planning, and to new domains such as indoor navigation and space station construction in a weightless environment.
Developing intelligent agents for human-computer interaction in time-critical domains such as radar operation, flight control, and robot teleoperation, using cognitive models of the user and task for plan recognition, advice generation, and information retrieval in anticipation of user needs. Currently developing models and agents for assistance in satellite operations.
Lecturer for Programming Robots senior level course in the Computer Science Department. Responsible for course content, structure, and execution. Created a multi-disciplinary project-focused course, in conjunction with a Mechanical Engineering professor, in which teams of students from the Computer Science, Mechanical Engineering, and Electrical Engineering Departments design, build, and program mobile robots that will compete in a jousting tournament; the winning robot will be entered into the Japan Robot Grand Prix.
Developed a map-based planner and a behavioral control layer that, combined with perception, were successfully used to navigate the DARPA Autonomous Land Vehicle in outdoor terrain. Designed and implemented a connectionist behavior-based architecture for mobile robot control that allowed for shared rather than traded control and was used for sonar-based indoor navigation. Collaborated in development of Gradient Fields algorithm for map-based planning, and incorporated it into the connectionist architecture. Developed simulation of a vehicle with range sensors in an outdoor environment which I then used to experimentally verify the hybrid control architecture. Also collaborated on the design and implementation of an adaptive component of the architecture that was used to provide fault-tolerance in autonomous underwater vehicles. Participated in writing of proposals for numerous projects as well as in personnel hiring process.
In collaboration with a cognitive scientist, designed a model of task-user interaction for a radar operator tracking and monitoring ships and planes aboard a navy vessel. Created a cognitively plausible model of the radar operator and the task as a basis for implementing an intelligent agent. Specified the model in a language designed to translate the model into a functioning expert system. Simulated the operation of the model to verify its correctness and predictive capability. Designed intelligent agent for assisting computer users by anticipating task demands, prioritizing and filtering information, and ensuring consistency and thoroughness in task execution. Applied for external grant to continue research independently.
Developed an interactive map-based mission planning system for mobile robots; designed control station for multiple robots; designed and simulated a real-time expert system for a developmental computer architecture; served as technology transfer liaison for Artificial Intelligence Center.
Served on the Program Committee for ROBOLEARN-97: Evaluating Robot Learning, held in conjunction with the 10th International Conference of the Florida Artificial Intelligence Research Society (FLAIRS-97).
Back to top
Back to Julio Rosenblatt's home