Autonomous Driving SoC Architecture for Redundant Sensor Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving systems fail to provide Level 4 or Level 5 functionality while meeting industry safety standards, as they require a dedicated supercomputer that is energy-efficient and low-power to process vast amounts of data from sensors in real-time, and achieve ASIL D functional safety without human intervention.
Innovation Solution
An end-to-end platform with a flexible architecture that leverages computer vision and ADAS techniques for diversity and redundancy, combined with deep learning tools, and a faster, more reliable, and energy-efficient System-on-a-Chip (SoC) integrated into a flexible, expandable platform for autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a dedicated supercomputer is used to process vast amounts of sensor data in real-time, then autonomous driving functionality and reliability are improved, but energy consumption and power requirements increase
Solution Approach 1:
The system segments processing tasks across multiple specialized processors including CPU, GPU, FPGA, and ASIC components, each handling specific aspects of sensor data processing. This segmentation allows efficient utilization of computational resources while reducing overall energy consumption compared to a single dedicated supercomputer.
Solution Approach 2:
The patent employs parameter changes by dynamically adjusting processing precision and computational intensity based on driving conditions. The system varies processing parameters to match actual needs, reducing energy consumption during normal operation while maintaining high reliability when needed.
2Reliability
If diverse sensors and processing systems are integrated for Level 4-5 autonomous driving, then functional safety and redundancy are improved, but system complexity increases
Solution Approach 1:
The system divides complex autonomous driving functionality into separate processing domains handled by different processor types (CPU for control, GPU for vision, FPGA for real-time processing, ASIC for specialized functions). This segmentation manages complexity by organizing diverse sensors and processing systems into modular, manageable components.
Solution Approach 2:
The patent implements a unified platform architecture where multiple processor types share common interfaces, memory systems, and control mechanisms. This universality allows the system to handle diverse sensor inputs and processing requirements through a standardized framework, reducing overall system complexity.
3Measurement precision
If real-time processing of sensor data is performed with high accuracy, then autonomous driving safety is improved, but computational power requirements and energy consumption increase
Solution Approach 1:
The system applies different processing precision levels to different sensor data streams based on their specific requirements. Critical safety-related data receives high-precision processing, while less critical data uses optimized lower-precision processing, reducing overall computational power requirements while maintaining necessary safety accuracy.
Solution Approach 2:
The patent replaces general-purpose mechanical computing with specialized hardware accelerators including GPU for parallel processing, FPGA for reconfigurable logic, and ASIC for fixed-function optimization. This substitution provides high processing accuracy with reduced power consumption compared to traditional CPU-based systems.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Autonomous driving is one of the world's most challenging computational problems. Very large amounts of data from cameras, RADARs, LIDARs, and HD-Maps must be processed to generate commands to control the car safely and comfortably in real-time. This challenging task requires a dedicated supercomputer that is energy-efficient and low-power, complex high-performance software, and breakthroughs in deep learning Al algorithms. To meet this task, the present technology provides advanced systems and methods that facilitate autonomous driving functionality, including a platform for autonomous driving Levels 3, 4, and/or 5. In preferred embodiments, the technology provides an end-to-end platform with a flexible architecture, including an architecture for autonomous vehicles that leverages computer vision and known ADAS techniques, providing diversity and redundancy, and meeting functional safety standards. The technology provides for a faster, more reliable, safer, energy-efficient and space- efficient System-on-a-Chip, which may be integrated into a flexible, expandable platform that enables a wide-range of autonomous vehicles, including cars, taxis, trucks, and buses, as well as watercraft and aircraft.