The Newton’s Polynomial Based - Automatic Model Generation (AMG) for Sensor Calibration to Improve the Performance of the Low-Cost Ultrasonic Range Finder (HC-SR04)

Abstract

The ultrasonic range finder sensors is a general-purpose sensor to measure the distance contactless. This sensor categorized as low-cost sensor that widely used in various application. This sensor has a significant deviation that lead to significant error in the measurement result. The error that produced by this sensor tends to increase proportionally to the measured distance. The implementation of the particular algorithm is required to reduce the error value. The model-based calibration is a solution to increase the accuracy. The model-based solutions are no longer feasible if the states of the model have changed. The longer of the usage of the sensor lead to sensor fatigue. Sensor fatigue is one of the causes of model state changes. As long as the drift still within the tolerance limit, the performance of the sensor still can be restored by using calibration method. The model-based calibration calibrates the sensor by using the model. The update of the model must be made whenever the changing of the model state occurred. Since the manual model making process is not an easy task, time and cost required, then the Newton polynomial-based AMG (Automatic Model Generation) have been implemented to this research. The AMG algorithm generates the new sensor model automatically based on the most updated states. This automatic model generation is implemented in the calibration process of the ultrasonic sensor. The implementation of polynomial-based AMG algorithm for sensor calibration have been succeeded to improve the accuracy of the calibrated sensor by 96.4% and reduce the MSE level from 25.6 to 0.914.

The reduction of polynomial degrees using moving average filter and derivative approach to decrease the computational load in polynomial classifiers

“…The more horizontal line elements increase the oscillation between the edge of data points of the horizontal line. The oscillation leads to a decreased accuracy level of the model result [23]. For both reasons regarding horizontal line elements, therefore horizontal line element removal is required.…”

“…The result of the feature vector that has been processed must be confirmed regarding the occurrence of Runge's phenomenon. The verification process involves the AMG kernel engine that was proposed by Gandha [23]. This kernel engine automatically generates a polynomial model based on a set of data points.…”

“…The discussion regarding the AMG engine kernel is optional. This research adopts a similar polynomial AMG kernel engine that was proposed by Gandha [23].…”

“…The number of data points is categorized as large data points. Without the degree reduction method, this condition leads to the occurrence of Runge's phenomenon when modeling using polynomial-based Automatic Model Generator (AMG) was done [23]. The AMG would generate a polynomial model with an n − 1 degree.…”

“…The number of data points affected the degree number of the generated model. The degree of the polynomial model would be generated for data points [23]. The unnecessary data point increases the degree number of the model.…”