Enhancing driving cycle development using artificial intelligence

In recent years, artificial intelligence (AI) has found application in numerous technical areas, including the automotive research and development sector. This paper considers the use of AI tools for the development of driving cycles for testing vehicles on a chassis dynamometer. The above idea was investigated on the example of a driving cycle simulating the use of a passenger car in urban conditions. The empirical data were collected during vehicle road tests in real traffic and then processed statistically by determining the values of selected driving pattern characteristics. Sections of vehicle velocity courses (‘micro-trips’) were selected and combined into a driving cycle representative of the road conditions prevailing during road tests. Processing of empirical data and combining velocity sections into a driving cycle was performed using AI-enhanced software utilizing large language models that convert user commands in natural language into Python code. The developed driving cycle was compared with selected standard urban driving cycles in terms of the values of driving pattern characteristics.