This research had been developed a system mainly consists of Arduino microcontoller based hardware and neural network based algorithms. The system has been fully assembled and successfully tested. By using two different methods the point feature detector (PFD) method was used as the first method. An Eigen Feature function was utilized to detect feature point of image. The second method is convolutional neural network (CNN) to recognize human face. Using PFD method, a classification value has been setup <11. The classification value is used as classification category of the program to recognize the subject (face image) correctly. By using PFD method, the response of the system from starting of a face image recognition until opening the locker is 20 second. The CNN method used alexnet to classify the image. At least around 300 training input data are use per person. The face recognition’s experiment reached a high recognition’s accuracy of 99.99% level and an average response time of 10 seconds. This research presents how the human face can be recognized and used to control the opening of a door lock.

Indoor test on the tire using the drum test machine by applying pressure and rotation according to the test method that has been determined until the tire is damaged, even until the tire explodes. This is done by measuring the maximum strength of the tire. Damage or explosion on the tire when testing takes place is not expected to occur, because this explosion can damage the drum test machine. For research and development purposes it is necessary to find out the location of the initial damage to the tire so that it can be improved on the weak part of the tire. A few moments before the tire is damaged a lump on the tire occurs and produces an unusual sound. This change in the frequency of sound that occurs before and when the initial tire is damaged will be analyzed, which will then be used to create a tire damage detection system so that the test can be stopped when the tire has initial damage.

Greek yogurt production needs a straining process that takes 10 hours or more. This paper proposes automation and control method for the centrifugation system to speed up the process time and to optimize the accuracy of quantity of whey drainage. Using system identification, the estimated mathematical model of straining process has been developed based on the traditional process of straining the yogurt. Then, the simulation and control design optimization has been carried out by using the estimated mathematical model. Based on the simulation results using whey mass controller, motor speed controller, and the combination of whey mass and motor speed controller, the controller that used are PID controller and fuzzy logic controller. The fastest controller is a PID controller as motor speed controller and fuzzy logic controller as whey mass controller that can speed up the production time and optimize the accuracy of quantity of whey drainage.

The winding process is widely used in manufacturing industries. For the high speed winding, the centre winding method is used. Irregular internal stresses at the centre of the roll result in major weaknesses such as buckling, spoking and cinching. Therefore, entanglement with the right tension is very important to get a stable wound package. It should be mentioned that winders usually operate based on the principle of precision winding. A typical characteristic of winders is increasing the surface speed as the diameter increases. This will cause the winding to increase which has the potential to cause damage. To overcome damage due to an increase in roll tension, the rolling motor speed must decrease so that the roll tension remains the same or even decreases. And because the rolling process uses additional media in the form of a liner fabric which is tension controlled with pneumatic disc brake, the pressure brake must also be made taper constructing both.