|
Plenary Lecture
Safety, Uncertainty, and Real-Time Problems in Developing Autonomous Robots
Professor Vitaliy Rybak
Department of Postgraduate Studies
Tecnological University of the Mixteca
Carretera a Acatlima Km. 2.5, Huajuapan de Leon, Oaxaca
MEXICO
E-mail: rybak@mixteco.utm.mx
Abstract: Recent
developments in robotics outside traditional industrial applications
increasingly focus on operation of robots in an unstructured environment and
human vs. robot interactions. Examples of new applications of robots in
unstructured environments that are actively pursued today are personal and
service robotics, space and underwater robotics, medical and rehabilitation
robotics, construction robotics, and agriculture robotics. The new trends in
robotics research have a general goal of getting robots closer to human
social needs. In this case a key problem of robotics is the problem of
safety of robot and its surrounding. For safe autonomous functioning in a
dynamic unstructured environment, a robot should possess a capability of
real-time data processing under information uncertainty. These three issues
- safety, uncertainty, and real-time data processing are closely related:
planning safe actions based on uncertain data usually requires more
computation than planning without uncertainty because multiple possible
outcomes of actions should be considered. The main sources of the
uncertainty are inaccuracy of sensorial data measurement, time-delay of
sensorial data acquisition and processing, and time-delayed feedback in a
robot’s control system. As increase of accuracy of measurements and speed of
data processing has technical and economic restrictions, there is a
necessity for search of practical decisions in the conditions of existing
possibilities. We propose a method of real-time data processing under
information uncertainty that deals with different acceptable levels of
uncertainty and collaborative processing of various data, visual and
non-visual. Finally we present the examples of application of the proposed
method for building of robot’s environment model and for robot motion
planning with obstacles avoiding.
Brief Biography of the Speaker:
Vitaliy Rybak received the Diploma in radio-physics from the Kiev State
University, Ukraine, 1958, and Ph. D. degree in Technical Cybernetics from
the Institute of Cybernetics of the Academy of Sciences of Ukraine, 1968.
Since 1958 till 2000 he was with the Institute of Cybernetics of the Academy
of Sciences of Ukraine. From 1975 to 2000 he was the scientific director of
the National Scientific Seminar of Ukraine “Scientific and Engineering
Problems of Robotics”. From 1982 to 2000 he was the head of Department of
Informatics in Robotics of the Institute of Cybernetics of the Academy of
Sciences of Ukraine. From 1989 to 2000 he was the director of the Section of
Robotics of the Scientific Council of the Automation of the National Academy
of Sciences of Ukraine. From 2000 till now he is a professor of the
Technological University of the Mixteca, Mexico; director of the Laboratory
of Robotics of the same university.
His major research interests include Intelligent Robotics (autonomous robot
architecture, 3D robotics vision, 3D stereo measurement, 3D object
recognition and scene analysis, goal directed robot’s behavior planning),
Image Processing, and Pattern Recognition.
He has published the book and more than 140 papers in Intelligent Robotics,
Image Processing, and Pattern Recognition. He was the responsible editor of
13 books in Artificial Intelligence, Intelligent Robotics, Image Processing,
and Pattern Recognition. He was the director of numerous international and
national research projects in Image Processing, Pattern Recognition, and
Intelligent Robotics.
He is the winner of the National award in the field of science and
technology of Ukraine, 1993.
| 
