Prediction of lithium-ion battery service life using data-driven methodologies in autonomous vehicles

Autonomous vehicles (AVs) rely heavily on lithium-ion batteries because of their efficiency, dependability, and high energy density. An accurate service life estimate is crucial for vehicle safety, performance optimization, and sustainability. However, the conventional opinion is that standard operating procedures fail to consider all the environmental and operational variables contributing to battery failure. This research aims to create an accurate and efficient forecasting model for autonomous vehicle lithium-ion batteries' estimated remaining service life. The suggested solution hopes to overcome shortcomings in existing approaches by using cutting-edge data-driven strategies. Using charge–discharge cycles, temperature profiles, and degradation measures as inputs, a new Service Predictive Data Classification Technique (SPDCT) is presented. This automated system gathers battery performance data throughout the operatival and idle phases. The data were analyzed using random forest classification in two periods. The initial classification analyzes operational-hours data to establish routine servicing intervals, whereas the subsequent classification concentrates on identifying abrupt battery spikes during operation that may necessitate prompt intervention to avert vehicle malfunctions. The categorization method, utilizing past battery operation and maintenance data, allows precise fault prediction, optimal service scheduling identification, and the prevention of operational failures. This methodical methodology guarantees ongoing surveillance of battery health metrics, allowing prompt maintenance actions as required. Including SPDCT substantially diminishes the likelihood of unforeseen battery problems during autonomous operations. The technology adjusts to various battery consumption patterns and climatic circumstances, offering tailored maintenance regimens for each vehicle. The extensive battery management technique using SPDCT improves the autonomous vehicle's performance, dependability, and operational safety, representing a notable progression in AV technology. Operational stress elements, including fast acceleration and large current demands, were used to increase the forecast accuracy further. These findings highlight how crucial it is to use multidimensional data for accurate performance prediction.