Implementation of K-NN Fingerprint Method on Receiving Server for Indoor Mobile Object Tracking

Tracking system is a way to make estimating the location of an object easier. One of this system ways uses to track the trolley position in the airport. This way is used because the passengers are difficult to find the trolley especially during busy time. One of study case from indoor mobile object is the movement of trolley inside the airport. This research used K-NN Fingerprint method on the receiving server to track the trolley position in Airport. The Fingerprint method is used to collect offline data from the received strong bluetooth signal. The K-NN method is used to classify fingerprint result of strong signal data because it is considered simple and has high speed in classification process. Information about the trolley position can be used to return the trolley to the trolley pickup. So, the passengers can easily get the trolley. From the tests carried out using the K-NN Fingerprint method on system, it was found that on the single object, MSE with the smallest result was when using k = 1 which is 0.6866 meters. In multi objects, MSE with the smallest result is when using k = 1 which is 3.3951 meters and tended to increase to the highest when using k = 10 which is 13.1432 meters. Meanwhile, the largest number of grids that is 54,825 grids produces the smallest MSE and the number of grids that is at least 3,531 grids produces the highest MSE. For the computational time test of the system, changes in the value of k have no affect on the time produced, while changes in the number of objects and changes in the number of grids on the Fingerprint affect the time produced.