Maximizing vehicle energy efficiency and its performance is a high priority for automotive industries as customers’ expectations rise. Engineers constantly face the challenge of balancing the conflicting goals of achieving superior performance and maximizing energy efficiency, all while meeting increasingly tight development timelines. Leveraging digital methods can potentially enable considerable reduction in development timelines.
Driving cycles function as standardized measurement procedures for certifying vehicle fuel efficiency and driving range. Representative velocity profiles condense numerous real-life driving cycles to enable quicker energy analysis and driver feedback evaluations. This paper introduces a novel methodology for generating synthetic drive cycles, such as average velocity cycles and ideal consumption velocity cycles, based on real-life driving scenarios.
In this study, the importance of creating representative drive cycles to enhance vehicle performance and energy consumption is highlighted. A detailed analysis of city cycle drives is provided, focusing on the prediction of average and ideal consumption velocity profiles using advanced machine learning, deep learning, and reinforcement learning techniques available in the field of data science. By leveraging data from numerous drives conducted at various ambient temperatures throughout the year, these velocity profiles were forecasted. The validation of these models is discussed in the paper, along with how effective and usable this method can be to save time in the overall vehicle development process. The merits and demerits of each ML model used to create synthetic drive cycles are discussed in the paper. This research offers valuable insights for automotive testing and development, significantly improving driver feedback analysis and providing essential insights to enhance driving habits and overall vehicle efficiency.