Automotive Power Window Communication with DTC Algorithm and Hardware-in-the Loop Testing

• Published in: Wireless personal communications Vol. 114; no. 4; pp. 3351 - 3366
• Main Authors: Kumar, Roushan, Ahuja, Neelu Jyoti, Saxena, Mukesh, Kumar, Adesh
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.10.2020; Springer Nature B.V
• Subjects: Algorithms; Automobiles; Communications Engineering; Computer Communication Networks; Controllers; D C motors; Decision trees; Electric motors; Engineering; Hardware-in-the-loop simulation; Machine learning; Networks; Position sensing; Signal,Image and Speech Processing; Temperature sensors; dSPACE; Force sensor; Decision tree classifier (DTC) algorithm; Power window; Current sensor; Hardware-in-the-loop
• ISSN: 0929-6212; 1572-834X
• DOI: 10.1007/s11277-020-07535-4

The power window control is the mechanism to automatically control the upward and downward movement of power window in vehicles by application of position, current, flexi force and temperature sensor as replacement of conventional or mechanical controlled hand crank based system. The paper focuses on model based development and testing of miniaturized scale controller for automotive smart power window, hardware-in-the-loop (HIL) testing using dSPACE simulator. The experimental work is carried out for the same to provide the desired precision due to the testing limitation in real time vehicle. The HIL set up is utilized to analyze the different manual and sensory inputs functionalities in Indian driving environment. The paper presented the DC motor interfacing to the controller as the main building block of the power window system. The motor takes the decision based on the decision tree classifier (DTC) machine learning algorithm and control movements in desired direction by condition inputs from decision tree. The system behavior is completely based on the algorithm, dSPACE testing environment supports the window functionality and its movement control in upward, downward direction and the results showed that the developed system is effective in identifying obstacle. The decision tree test set and training set data depicts that 93%-sampled results are valid outcome and 7% are invalid under 12 different test observations.