<![CDATA[Fires often occur in the area of housing, Office space or in the Woods. The emergence of the fires themselves often leave casualties is not a little. Fire occurrence process itself can develop quickly or slowly depending on various factors such as the typically temperature, wind direction or weather based on material to burn. The issue was made of various fire alarm systems that are usually placed on building large and function gives a warning if the fire is large enough or usually the alarm will also be activated manually but the system does not tell you the location of the occurrence of fires. Based on the problems researchers make fire with point detection system sensor LM35 and flame sensor using algorithms K-Nearest Neighbor and a microcontroller, Arduino Mega as data processing. The workings of this system with sensor LM35 temperature detected on duty around the room and the sensor detects flame flame candles. If there is a fire then the system will process the calculation by the method of classification results obtained so KNN are used as the end result of this system. After the research is done, the results obtained from testing system has the accuracy of 80.55% and for process calculation of classification are obtained for 1428.83 ms.]]>
Copyrights © 2019
