With the development of automotive intelligence, the role of cars as robots with artificial intelligence is particularly obvious. General-purpose processing chips can no longer meet the needs of automobiles: multi-sensor fusion, high data throughput, high bandwidth, low latency, parallel computing of multi-core DSP, deep learning capabilities and cloud service acceleration capabilities all require chips with new architectures. Chips for automotive electronics are far more reliable and environmentally adaptable than ordinary consumer and industrial-grade chips can provide.
ADAS Leader has accumulated many years of chip design experience, designed chips for ADAS systems, and gradually invested in research and development, providing powerful computing power, cloud service capabilities, and highly reliable chip systems for future autonomous driving systems.
4 X MIPS32
4 X VMP
Main frequency dual-core 1G
Dedicated vector processing unit
ARM+GPU+Sensor HUB
The floating-point operation speed reaches 12.5 trillion times per second; the DOS system bandwidth can support 40Gbps; support 3D scene restoration macro understanding recognition; (with macro understanding cognitive ability) support 20 high-resolution sensors at the same time; comply with ASILB(D), IS026262 specifications: power consumption within 5w, passive heat dissipation design: 7nm FINFET process.
A1+DOS
True fusion neural network sensor management with understanding cognitive ability (increasing microscopic understanding cognitive ability), central processing unit of scheduling system, brain system integrator.
A2+TPU+5G