Biological Systems Engineering

 

Date of this Version

2016

Citation

Published in Computers and Electronics in Agriculture 128 (2016), pp 141–148. doi:10.1016/j.compag.2016.09.001

Comments

Copyright © 2016 Elsevier B.V. Used by permission.

Abstract

Various hardware and software solutions exist for collecting Controller Area Network (CAN) bus data. Digital data accuracy could vary based upon different data logging methods (e.g., hardware/software timing, processor timing, etc.). CAN bus data were collected from agricultural tractors using multiple data acquisition solutions to quantify differences among collection methods and demonstrate potential data accumulation rates. Two types of data were observed for this study. The first, CAN bus frame data, represents data collected for each line of hex data sent from an ECU. One issue with frame data is the resulting large file sizes, therefore a second logging format collected was an averaged frame signal, or waveform dataset. Because of its smaller file size, waveform data could be more desirable for long periods of collection. Percent difference was calculated from two sets of frame data logs using different hardware/software combinations, and a frame data log was also compared to a waveform data log. The resulting difference was less than 0.0025 RPM for engine speed comparisons, zero for fuel rate and fuel temperature comparisons, and the mean percent difference was less than 0.08% between the methods of data collection. The error production could have resulted from noise in hardware and processor times, but was not found to increase as time progressed. This showed that even though errors existed between logging methods, the magnitude of errors would not negatively impact any practical agricultural field research applications. Thus, data logged by the different devices was similar and files requiring less memory would be desired. Selecting a waveform CAN bus data logging option would likely maintain digital data accuracy while reducing file storage and processing needs.