RE: Occula27 Oct 2021 17:27
Cont...
1. Work exclusively with an automotive-grade semiconductor supplier to create an Occula #ASIC chip. Using Mobileye as a guide, an obvious candidate would be STMicroelectronics to act as foundry, with Seeing Machines the owner of the chip and all IP. This would mirror the original Mobileye/ST #EyeQ chip partnership precisely.
2. Work with multiple semiconductor chip vendors who each individually develop an ASIC with the Occula NPU embedded. This would mirror the Arm Holdings business model of #IP licensing to multiple silicon partners and could quickly lead to a defacto DMS standard, as seen with Arm Cortex-A/R/M cores. Examples could be: Qualcomm fifth-generation ****pit applications processor, Mobileye EyeQ6, NVIDIA Atlan, Texas Instruments Sitara, or even ASICs/SoCs developed in-house by the automakers themselves (many rumors abound).
What is clear is the demands of the DMS processor are rising exponentially, including first distraction and drowsiness, now moving to include impairment (first alcohol, then drugs) with a transition also from DMS to OMS (full-cabin) monitoring.
The benefits in terms of performance and power consumption of a software/hardware co-design approach and having total ownership of both the algorithms and silicon processor will already be abundantly clear to automakers as they make comparisons with a hardware agnostic design approach and plan a DMS/OMS feature roadmap for the next decade.
Software/hardware co-design is the same path as followed by Apple with the development of the new M1 Pro and M1 Max processors, discussed here: https://lnkd.in/dNVqtKuK
It is extremely unusual to see this level of system design competency in an approx. 200 person company.
Sign In
Follow the stocks
