BYD’s intelligent driving training data is insufficient, seeking solutions

According to a person in charge of BYD's intelligent driving research and development, BYD's intelligent driving training data has two major problems: first, the data volume is large, but the data is highly repeated and not rich enough to cover all possible scenarios; second, the data quality is not high. Many old cars are technically limited, and even some high-end models cannot send back high-quality driving data. To solve this problem, BYD has begun to study world models to improve the quality of simulation data.