Dual-Processing Unit Architecture for ADAS Image Stream Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Camera-based Advanced Driver Assistance Systems (ADAS) face challenges in reducing memory and computing requirements while maintaining safety features under varying light conditions, leading to increased costs and complexity.
Innovation Solution
The implementation of a dual-processing unit architecture that generates two data streams with reduced bandwidth, allowing for efficient image pre-processing with lower memory demands, utilizing separate processing paths and shared memory to optimize RAM usage and reduce data rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high dynamic sensors are used to provide desired safety under varying light conditions, then safety performance is improved, but memory and computing requirements increase
Solution Approach 1:
The image data stream is divided into multiple processing paths (first processing unit and second processing unit), each handling different aspects of image processing. This segmentation allows the system to process high-resolution images for safety-critical functions while using reduced-bandwidth streams for less demanding tasks, thereby reducing overall memory requirements while maintaining safety performance.
Solution Approach 2:
Different processing paths are assigned different quality levels. The first processing unit generates a first data stream with reduced bandwidth suitable for certain processing tasks, while the second processing unit generates a second data stream with different characteristics. This local quality differentiation allows the system to use high-quality processing only where necessary for safety, reducing overall memory and computing requirements.
2Reliability
If high dynamic sensors are used to provide desired safety under varying light conditions, then safety performance is improved, but computing performance requirements increase
Solution Approach 1:
The computing workload is segmented across multiple processing units with different processing capabilities. By dividing the image stream into multiple paths with different bandwidth requirements, the system distributes computing tasks appropriately, avoiding the need for excessive computing power in a single processing unit while maintaining the ability to handle high dynamic range images for safety functions.
Solution Approach 2:
Different processing paths are configured with different computational intensities matched to their specific tasks. The first processing unit operates with reduced bandwidth for tasks that don't require full image quality, reducing local computing requirements. This allows the system to maintain high safety performance through selective high-quality processing only where absolutely necessary.
3Reliability
If camera-based systems are equipped with high dynamic sensors, then safety performance is improved, but system cost increases
Solution Approach 1:
The system segments the image processing into multiple paths, allowing the use of high dynamic sensors only where absolutely necessary for safety-critical processing. Other processing paths can use reduced-bandwidth streams that require less sophisticated (and less expensive) processing hardware, thereby reducing overall system cost while maintaining safety performance.
Solution Approach 2:
The system applies different quality levels of processing to different data streams. By generating reduced-bandwidth data streams for certain processing tasks, the system can use less expensive processing hardware for those tasks while reserving high-quality processing only for safety-critical functions, thus reducing overall system cost while maintaining required safety performance.
4Measurement precision
If high-resolution image streams are processed, then image quality is maintained, but memory requirements increase
Solution Approach 1:
The image data stream is segmented into multiple processing paths. The first processing unit generates a first data stream with reduced bandwidth that can be stored in memory with lower requirements. The second processing unit handles the full-resolution data for processing tasks that absolutely require high image quality, minimizing the amount of high-resolution data that needs to be stored in memory at any given time.
Solution Approach 2:
The system maintains high image quality locally in the second processing unit where it is absolutely necessary, while the first processing unit operates with reduced bandwidth for tasks that don't require full image quality. This local quality differentiation allows the system to maintain overall image quality while significantly reducing memory requirements for storing and processing image data.
Data Source
AI summary
An implementation relates to a device for processing an image data stream. The device may include a first processing unit and a second processing unit for receiving the image data stream. The first processing unit may be arranged for providing a first data stream, the first data stream has a reduced bandwidth compared to the image data stream. The second processing unit may arranged for providing a second data stream, the second data stream has a reduced bandwidth compared to the image data stream.