Pattern Recognition Using a Neural Network on a Microcontroller with I2C Ultrasonic Sensors
Annals of Emerging Technologies in Computing (AETiC), Print ISSN: 2516-0281, Online ISSN: 2516-029X, pp 9-19, Vol. 3, No. 1, 1st January 2019, Published by International Association of Educators and Researchers (IAER), DOI: 10.33166/AETiC.2019.01.002
11 Pages Posted: 23 Feb 2020
Date Written: January 1, 2019
Ultrasonic sensors have been used in a variety of applications to measure ranges to objects. Hand gestures via ultrasonic sensors form unique motion patterns for controls. In this research, patterns formed by placing a set of objects in a grid of cells are used for control purposes. A neural network algorithm is implemented on a microcontroller which takes in range signals as inputs read from ultrasonic sensors and classifies them in one of four classes. The neural network is then trained to classify patterns based on objects’ locations in real-time. The testing of the neural network for pattern recognition is performed on a testbed consisting of Inter-Integrated Circuit (I2C) ultrasonic sensors and a microcontroller. The performance of the proposed model is presented and it is observed the model is highly scalable, accurate, robust and reliable for applications requiring high accuracy such as in robotics and artificial intelligence.
Keywords: Ultrasonic Sensors; Time of Flight; Echo; Transmitter; Receiver; I2C; Neural Network; Pattern Recognition; SONAR
Suggested Citation: Suggested Citation