8b4e529225152b2fa6749c10c1d616dd-Paper-Conference.pdf

In autonomous driving, designing a single reward often requires reliance on complex parameters and hard-coded knowledge, making the agent less efficient and adaptive(Kamranetal.,2022).