In the current intelligent upgrade of automobiles, the traditional distributed electronic and electrical architecture fails to meet the requirements of autonomous driving due to insufficient computing power, difficulty in real-time information transmission, and non-compliance with security standards. The current intelligent vehicle architecture, through the integration of central computing and regional control, combines heterogeneous computing units and security systems to form an "in-vehicle brain", providing support for the implementation of autonomous driving. This article studies the key points and core components of the intelligent vehicle computing platform technology, clarifies the three major technical directions of architecture design, communication technology, functional safety and energy efficiency optimization, analyzes the application characteristics of heterogeneous computing chipsets and key auxiliary devices, and provides references for the improvement of platform performance and industrial application.