Target detection using intelligent robots with continuously syncronized sensors

The typical properties of sensors used on robot affect the capability of the robot directly. A robot receives data from the surrounding area. The accuracy of these data are more important. More than one sensor are used to get more accurate data and form the sensor integration. The sensors may have same or different characteristic properties according to each other in integration. The characteristic properties must be taken in consideration in sensor integration to analyse data. Different measuring boundaries and response times of distance measuring sensors? integration is focused in this thesis. If the value to be measured in an integration composed of the sensors is out of ranges of some sensors, the data of these sensors is to be eliminated. The methods used Gauss probability are insufficient if the boundaries are not taken in consideration. Chauvenet?s or Peirce?s criterions are bad data elimination methods. Logical output of sensors are used in Chauvenet?s criterion mathematical model to increase reliability in this thesis. To increase reliability once more measuring repeated for each sensor in an integration. Also, lots of robots can be used to determine and analyze more options of one target or targets. The aim of this thesis, in the presence of the criteria, is to develop new methods to detect the target accurately by the sensor integration on intelligent robots. Computer network, statistical analyzing methods, logic gates, software technologies are used in this work. RoboKS2 is designed, to direct some studies, apply and test the new developed methods practically and help to create new ideas. Keywords: Chauvenet?s criterion, sensor integration, sensors with different meausuring boundaries and response time, syncronously data analyzing, error data analyzing, Gaussian distribution.