Pengenalan Individu Berdasarkan Gait Menggunakan Sensor Giroskop
Abstract: Every persons have
their own unique way of walking which is called gait. Gait can be used to
identify a person. Gyroscope is a sensor used to detect vibration and measure
acceleration based on direction or orientation. This paper presents an
individual recognition based on gait using gyroscope sensor embedded in
smartphone. The gait data is processed and analyzed by implementing Linear
Predictive Coding (LPC) and k-Nearest Neighbour (k-NN). LPC is used to extract
features from gait data. It produces feature vector based on combination of
p-previous signal and takes only important value of the feature data. Then, k-NN
is used for classification, using some calculation methods such as Euclidean,
Cityblock, Cosine, and Correlation distance. The gait signals contain x,y,z
axis and the signal magnitude. In this paper x,y,z axis and the signal
magnitude are also combined to improve the accuracy. The highest accuracy of
99.58% is achieved using signal combination x-y-z-m. Overall, this person
detection system produces accuracy between 50% to 99.58%.
Penulis: Muhammad Satya Annas,
Achmad Rizal, Ratri Dwi Atmaja
Kode Jurnal: jptlisetrodd170479
