Excerpt from Computer Design April 1992 NASA EYES RANGE OF FUZZY CONTROL IDEAS IN SPACE ------------------------------------------------ NASA has recently reported encouraging research results in a number of areas using fuzzy logic. One of the most advanced projects is a controller for space shuttle proximity operations, i.e. maneuvering around or keeping position with respect to another object in space. Work has been progressing on a fuzzy- based translational controller which deals with the parameters of azimuth and angle and their respective rates of change, and the range and rate of change of range with respect to another object. NASA engineers developed natural language rules to run the controller and are testing it in a multi-vehicle simulation by substituting the fuzzy controller for the simulator's normal human inputs. The rule base was learned from the experience of human operators and the efficiency of the controller was tuned based on flight profiles recorded from actual missions and simulations. One of the main advantages in developing the fuzzy translational controller were that the engineers did not need to construct a detailed mathematical model of the system in advance. Performance was honed through simulation and experience. The results of simulations have been encouraging, especially in terms of fuel efficiency. In holding position with respect to a target, the fuzzy controller required significantly less acceleration (i.e. smaller increments of position change) than did the human controlled simulation. In overall maneuvers, the fuzzy controller has shown a 20% to 70% better fuel efficiency than the currently used digital auto pilot and the best simulation runs of human pilots. NASA is also exploring other applications of fuzzy control in space. Among the projects being considered are the use of inexpensive cameras for constant tracking of objects around the space station, fuzzy control can contribute to collision avoidance systems, robot arm control and traffic management. At the great distance of interplanetary space where it can take 20 minutes to send a signal and receive an answer, robotic systems will have to operate quasi-independently. A fuzzy controller on an unmanned Mars rover vehicle is expected to help the rover avoid obstacles and identify and collect soil samples based on imprecise sensor input and only partially known conditions. ------------------------------------------------------------ This is article is provided with permission from Computer Design. For subscription information to Computer Design, call Paul Westervelt at (913) 835-3161. Do not redistribute in an form (written or electronic) without permission from Computer Design. This is information is provided by Aptronix FuzzyNet 408-428-1883 Data USR V.32bis