Abstract: To address the issues of high concurrency processing and hard real-time scheduling response capability improvement in operating systems in autonomous driving collaborative control scenarios, a real-time scheduler based on improved least laxity first (LLF) strategy was developed. The design and integration of the proposed LLF scheduler based on RT Linux were completed through constructing scheduler data structures, enhancing the red-black-tree task management mechanism, and introducing new scheduler classes and key scheduling functions. Experimental results regarding the real-time performance of the operating system and the response capability of vehicle collaborative control tasks demonstrate that the proposed improved LLF scheduler outperforms the original scheduling strategy in terms of scheduling delay and system jitter, achieves low latency and precise scheduling control of periodic and event-driven real-time tasks, and provides a feasible solution for high concurrency and hard real-time multi task scheduling in collaborative control scenarios of intelligent electric vehicles.