Indentification of Beef in Beef and Chicken Experiments using Conducting Polymer Sensor Series and Kohonen Algorithm Method

Abstract

Chicken, and beef each have a distinctive aroma. Identification of Chicken and beef based on the aroma of the meat using an electronic nose. This electronic nose uses a series of sensors consisting of 6 (six) pieces and uses a Conducting Polymer. This polymer has a high resistance so it is widely used as an insulator. However, this resistance has a certain limit where the polymer surface will turn into carbon and conduct electric current if exposed to excessive electric charge. This research was conducted by taking samples of chicken and beef as test samples where these meats were placed in a closed container at room temperature. Data is taken alternately every day to find out the odor of each meat where on the first day data is taken from the odor of chicken, and on the second day data is taken from the odor of beef. This condition is done to ensure the freshness of each meat. This study uses a Neural Network (NN) as pattern recognition and ATMega16 microcontroller as data acquisition. Neural Network is trained using Kohonen. The sensor used is a Conducting Polymer sensor because of the nature of the Conducting Polymer where the output is a voltage generated due to changes in the polymer resistance resistance. A two-layer neural network consisting of six input nodes and three output neurons is trained using the Kohonen algorithm with the training process completed in 31 iterations. The test was carried out 30 times for each exposure to steam from the odor of chicken and beef which was carried out alternately. The percentage of success of the system is 100%.