Distributed AV Computing Architecture for Fault-Isolated Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Centralized computing architectures in autonomous vehicles are prone to failures that can propagate and compromise safe operation, as they rely on a single high-performance system, leading to increased risk and complexity in integrating diverse modules and subsystems.
Innovation Solution
A distributed computing system is implemented, distributing the complexity of autonomous vehicle operation across multiple computing systems, reducing the reliance on a single high-performance system and enhancing redundancy, thereby mitigating failure risks and improving safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If a centralized computing architecture is used in autonomous vehicles, then high-performance processing can be achieved, but system reliability deteriorates due to single-point failures and failure propagation
Solution Approach 1:
The patent divides the centralized computing system into multiple distributed computing nodes (first computing system, second computing system, third computing system). Each node performs specific processing tasks independently, eliminating single-point failures. The segmentation allows the system to maintain high computing power while improving reliability through redundancy and fault isolation.
Solution Approach 2:
Different computing nodes are assigned specialized functions: the first computing system handles sensor data processing, the second handles trajectory planning, and the third handles control commands. This local quality assignment optimizes processing efficiency for each function while maintaining overall system reliability through distributed architecture.
2Reliability
If a distributed computing system is implemented, then system reliability is improved through redundancy, but device complexity increases due to multiple computing systems
Solution Approach 1:
The patent merges multiple computing systems into a coordinated distributed architecture where each node handles specific tasks. By combining specialized processing units with defined communication protocols, the system achieves high reliability while managing complexity through functional integration rather than isolated components.
Solution Approach 2:
Each computing node is designed with multi-functionality, capable of performing sensor processing, trajectory planning, and control functions as needed. This universality reduces the need for dedicated specialized hardware for each function, thereby reducing overall device complexity while maintaining reliability through redundant capabilities.
3Reliability
If distributed computing is used, then failure risks are reduced through redundancy, but data transmission loads increase due to communication between multiple systems
Solution Approach 1:
The patent extracts and processes data locally at each computing node before transmission. The first computing system processes sensor data locally, the second processes trajectory data locally, minimizing the amount of raw data that needs to be transmitted across the network. This extraction approach reduces data transmission loads while maintaining the reliability benefits of distributed processing.
4Productivity
If multiple computing systems are deployed, then processing parallelism is achieved, but integration complexity increases for diverse modules and subsystems
Solution Approach 1:
The patent implements dynamic task allocation and data routing between computing nodes based on real-time system state and task priorities. This dynamic approach allows parallel processing of multiple functions while adapting to changing conditions, reducing integration complexity through flexible coordination rather than rigid fixed assignments.
Data Source
AI summary
Disclosed are distributed computing systems and methods for controlling multiple autonomous control modules and subsystems in an autonomous vehicle. In some aspects of the disclosed technology, a computing architecture for an autonomous vehicle includes distributing the complexity of autonomous vehicle operation, thereby avoiding the use of a single high-performance computing system and enabling off-the-shelf components to be use more readily and reducing system failure rates.