NewsChauffeur Monitoring System ASIC With Integrated AI Processing Unit

Chauffeur Monitoring System ASIC With Integrated AI Processing Unit

Category articles

The on-chip NPU is supported by TensorFlow, Caffe, MXNet as well as ONNX. Additionally, this ASIC embeds quad Arm Cortex A5 CPU cores with Neon modern technology for accelerated video clip encoding/decoding and on-chip video clip analytics algorithms, along with hardware for photo processing, video clip encoding and RGB/IR processing. Its high dynamic variety (HDR) handling capability enables the ASIC to accept input from RBG/IR photo sensors and also support top quality output for videos taken throughout the day or in the evening. The incorporated video clip encoder approves up to 5 megapixel records from OmniVision’s auto image sensing units, as well as outputs as much as 2K resolution video at 30 frames per second (fps).

The OAX8000 consumes simply 1 watt in typical problems. This assimilation likewise minimizes the board area for the engine control system (ECU). Boot-up time for the ASIC is considerably fast, which gets rid of any kind of hold-up between ignition as well as activation of the DMS cam. Furthermore, it supports safe boot functions to give cybersecurity.

Other applications include refining occupant detection formulas, such as differentiating an infant from a grocery bag and offering alerts when items are left in the automobile. In addition, this ASIC can be used in automobile video clip safety and security systems to perform features such as FaceID, as well as predetermined driver-comfort settings (e.g. seat placement) that are activated when the DMS scans the driver’s face.

The OAX8000 ASIC is AEC-Q100 Grade 2 certified and readily available now from OmniVision Technologies in a BGA196 plan.

Optimised for entry-level, stand-alone motorist tracking systems(DMS)is the OAX8000, an AI-enabled, auto application-specific incorporated circuit (ASIC). The OAX8000 uses a stacked-die architecture to give the DMS processor with on-chip DDR3 SDRAM memory (1 Gb). It additionally incorporates neural handling system(NPU) and also image signal processor(ISP), which provides specialized handling speeds of approximately 1.1 trillion operations per secondly for eye tracking formulas. These rapid processing speeds with 1K MAC of convolutional semantic network (CNN) velocity, in addition to integrated SDRAM, allow low power usage for DMS systems.

Michal Pukala
Electronics and Telecommunications engineer with Electro-energetics Master degree graduation. Lightning designer experienced engineer. Currently working in IT industry.