Agricultural tractor driving cycle extraction using artificial intelligence

Driving cycle assessment is one of the common methods to evaluate a vehicle’s real-world condition also monitor fuel consumption and emissions. The basic challenge in the extraction of the driving cycle is data analysis to develop and define the suitable behavior of the device. Clustering, classification, and recognition of driving patterns are important steps in the extraction of a suitable driving cycle. Generally, the accuracy of modeling and recognition of AI-based methods is indicated by more than 90% and other outputs comply with big data. Thus, in this research, we endeavored to evaluate the effect of using artificial intelligence on the driving cycle of off-road vehicles. The major part of off-road vehicles are agricultural vehicles such as tractors which are divided into three categories based on agriculture operations; light, heavy, and extra heavy. In addition, the procedure of agricultural operation is effective on fuel consumption, loading, and exhaust emissions. The results of this research showed that the use of conventional machine learning methods for clustering and classification can be used for any volume of features. However, with an increase in features, the complexity of region segmentation and the effect of farm management factors cause overtraining conditions in the learning algorithm and reduce the accuracy of the extracted driving cycle and prediction of driving behavior. Therefore, it is necessary to use advanced algorithms with deep learning capabilities. Therefore, extracting the intelligent driving cycle for agricultural tractors based on the type of agricultural operation with the help of artificial intelligence methods can reduce fuel consumption, pollution, and optimal farm management.