Sets of Beloning in Problems of Signals Classification
Nikolay P. Lepekha, Nikolay Fedorovich Kirichenko, Roman B. Kuc · Journal of Automation and Information Sciences · 2001
New mathematical tools for a signal classification and space objects recognition are considered by using special container sets proposed as some canonical geometric bodies in an attribute space to contain in a compact form the attribute points of learning sets. The tools are suitable for writing the condition of noncrossing for the container sets within the rules of pattern recognition. The results of the article represent the development of the ideas of object recognition from ultrasonic echoes by using an adaptive and mobile sonar sensors described by the authors in the previous articles.