Redundant Processing Units With External Voting for Functional Safety
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high computational effort and increased response times in synchronizing and comparing results across multiple independent computing entities in functionally safe applications, particularly in railroad technology, become unacceptably high when using Commercial-off-the-Shelf hardware, leading to inefficiencies in processing networks.
Innovation Solution
A method is implemented where redundant results from a plurality of computing entities are generated and compared externally within a voting system, while internal messages within processing units are not compared, allowing asynchronous processing and reducing the need for synchronization, thus optimizing computational effort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If results from multiple independent computing entities are compared externally within a voting system, then functional safety is maintained, but computational effort and response times increase unacceptably high
Solution Approach 1:
The patent segments the computing entities into independent processing units that each perform redundant calculations locally. By dividing the system into autonomous units that process data independently and only exchange final results, the computational overhead is reduced while maintaining functional safety through external voting on the segmented outputs.
Solution Approach 2:
The patent extracts the voting comparison operation from the internal processing of computing entities and places it externally. This allows the computing entities to perform their calculations independently without continuous synchronization overhead, and only the final results are subjected to external voting, thereby reducing computational effort while preserving reliability.
2Reliability
If results from multiple independent computing entities are compared externally within a voting system, then functional safety is maintained, but response times increase unacceptably high
Solution Approach 1:
By segmenting the system into independent processing units that operate autonomously, the patent eliminates the need for continuous synchronization between computing entities. Each unit processes data independently and produces results that are later compared externally, significantly reducing response times while maintaining functional safety through the voting mechanism.
Solution Approach 2:
The computing entities perform their calculations and prepare results in advance before the external voting comparison takes place. This preliminary independent processing allows the system to avoid time-consuming synchronization during critical processing phases, reducing overall response times while ensuring functional safety through subsequent external validation.
3Reliability
If strict synchronization is implemented between independent computing entities for voting, then voting reliability is ensured, but computational complexity and overhead increase significantly
Solution Approach 1:
The patent extracts the synchronization requirement from the internal operations of computing entities and relocates it to an external voting stage. This allows the computing entities to operate independently without complex synchronization protocols, reducing device complexity while maintaining voting reliability through external comparison of the independently generated results.
4Ease of manufacture
If Commercial-off-the-Shelf hardware is used for processing, then cost is reduced, but synchronization and comparison effort becomes unacceptably high
Solution Approach 1:
The patent segments the processing architecture into independent units that can be implemented using standard Commercial-off-the-Shelf hardware components. By dividing the system into autonomous processing units with clear interfaces, the patent enables the use of cost-effective COTS hardware while avoiding the need for complex custom synchronization infrastructure, thereby maintaining processing efficiency.
Data Source
AI summary
A method executes a technical process by executing application programs redundantly in a redundancy group containing a plurality N of computing entities and thereby calculate a plurality N of redundant results. A comparison of the messages containing the results is performed. When a majority of the messages contain identical results the majority of messages is used for further execution of the technical process. Processing units are formed from a plurality of computing entities being combined in each of the processing units, wherein none of these processing units contains a plurality of computing entities from the same redundancy group. The generated redundant messages are compared to determine if the messages are to be sent as external messages to computing entities, which do not belong to the processing units. No comparison is carried out for messages that are exchanged as internal messages between computing entities within a processing unit.