The development of a proprietary system that utilizes AI to predict the values of key tire characteristics has been announced by Yokohama Rubber.
Yokohama started to use the system this month and predicts that it will enable the company to conduct a large number of virtual experiments, resulting in the acceleration of new tire development, reduced development costs, and better performing products. The manufacturer also says that the system will make it easier for less-experienced engineers to develop new tire designs.
Developed under Yokohama’s AI utilization concept called HAICoLab (Humans and AI ColLaborate for digital innovation), the latest AI system is capable of predicting values for key tire characteristics based on data supplied by tire designers. This includes specifications-related data for tire design parameters such as physical properties of compounds, material-related data, structure and shape, and evaluation conditions.
Yokohama states that the new system reduces the deterioration of AI prediction accuracy, a problem that can occur during the tire design phase. As design parameters differ depending on a tire’s internal structure, separate databases are needed and used for AI learning according to the tire’s structural features. Narrowly composed learning data can sometimes reduce the accuracy of AI prediction; however, the prediction accuracy of Yokohama’s AI is improved by transferring AI prediction ability from related areas.
The tire maker is also utilizing an earlier AI system that was first developed in December 2020 alongside the latest AI solution. By using both systems in tandem, Yokohama hopes to enhance the development of a range of new tires.
Yokohama aims to acquire knowledge by creating and collecting data based on hypothetical conditions set by humans. The company then applies the AI system to predict, analyze and search for optimal results.