Safe Local Machine Control Using Cloud-Verified Control Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating control data for local machines, such as vehicles, face challenges in ensuring safe operation due to limited processing power and memory of local hardware, while relying on unreliable network connections for cloud computations that are not suitable for safety-critical functions.
Innovation Solution
A method that integrates cloud computations into the calculation of local machine behavior by outsourcing calculations to a cloud application, ensuring that control data are within safe parameters and delivered on time, using a local software application to check and verify the safety of cloud-generated data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If cloud computations are used to generate control data, then processing power and memory are improved, but network reliability deteriorates
Solution Approach 1:
The system segments the control data generation process into two parts: complex computations are performed in the cloud, while safety-critical verification and final control decisions remain local. This segmentation allows leveraging cloud processing power while maintaining local reliability for safety-critical functions.
Solution Approach 2:
The local software application acts as an intermediary between the cloud application and the actuators. It receives control data from the cloud, verifies safety constraints locally, and only executes verified control data. This intermediary role ensures that even if network communication fails or cloud data is compromised, the local machine maintains safe operation.
2Reliability
If all calculations are performed locally, then reliability is improved, but processing power and memory are insufficient
Solution Approach 1:
The system segments computational tasks by complexity and criticality. Non-critical computations are offloaded to the cloud, while critical safety verification remains local. This segmentation enables the use of limited local resources for essential functions while utilizing abundant cloud resources for complex calculations.
Solution Approach 2:
The local software performs partial verification of safety constraints rather than complete recalculation of all control parameters. This partial action approach allows the system to maintain reliability for critical safety functions while accepting cloud-generated control data for non-critical parameters, effectively working within limited local processing capabilities.
3Reliability
If control data are verified locally, then safety is improved, but processing time increases
Solution Approach 1:
The local verification process performs only essential safety checks on cloud-generated control data rather than complete recalculation. This partial verification approach maintains safety guarantees for critical parameters while minimizing additional processing time, allowing the system to meet real-time control requirements.
Solution Approach 2:
The cloud application performs preliminary computation of control data before transmission to the local machine. This preliminary action reduces the amount of verification needed locally, as the cloud has already performed extensive calculations. The local system only needs to verify safety constraints rather than perform full computations, reducing overall processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and system to generate safe control data to control and/or to supervise a local machine (L-MACH), wherein a local software application (L-APP) is executed on a safe hardware device of a safe device (S-DEV), which local software application (L-APP) executes calculations based input data to generate said safe control data, wherein a software application in the cloud, a so-called "cloud application" (C-APP), which is executed on a hardware external to the safe device (S-DEV), and wherein the local software application (L-APP) sends a request to the cloud application (C-APP), the request comprising a request to the at least one cloud application to execute one or more defined calculations and said specific input data, which is to be used by the at least one cloud application (C-APP) for the execution of the one or more defined calculations, and wherein if the local software application (L-APP) timely receives control data of requested calculations it is checked if said control data are within a safe parameter range, and if said control data are within the safe parameter range, the control data are qualified as safe control data and will be used to control and/or to supervise the local machine (L-MACH), and if said control data are not within the safe parameter range, said control data will not be qualified as safe control data and will not be used to control and/or to supervise the local machine (L-MACH), and wherein if the control data are not received timely they will not be used to control and/or supervise the local machine (L-MACH).