Abstract: Kinetic equations serve as fundamental mathematical models for describing the dynamic behavior of particle systems. Their intrinsic multiscale structure, however, poses significant challenges in developing efficient and robust numerical schemes. A popular approach is to design asymptotic preserving (AP) schemes. It aims to create a unified solver across the computational domain that captures the correct asymptotic limits at a discrete level. In recent years, we have developed a class of unconditionally stable AP schemes based on characteristic backtracking. These methods feature low computational complexity, high parallel efficiency, and excellent scalability, making them particularly suitable for long-time simulations of high-dimensional, multiscale transport phenomena. Owing to these advantages, the proposed framework holds great potential for applications in national defense engineering, astrophysics, semiconductor, and biochemistry.