Serial Data-Stream Dropout Mitigation with ML-Predicted Actions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Communications channels are prone to interruptions due to noise and interference, leading to undetected or erroneously detected data, which can have severe consequences in serial data-streams, such as loss of control in aircraft systems.
Innovation Solution
A system utilizing machine learning modules to predict and generate future data packets based on valid messages, substituting for corrupted data during interruptions, ensuring continuous operation by performing predicted actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If data is transmitted over communications channels, then information is communicated, but data dropout occurs due to noise and interference
Solution Approach 1:
The system performs preliminary actions by predicting multiple future operations ahead of time using machine learning models. When data dropout occurs, these pre-predicted operations are executed in sequence to maintain continuous control. This resolves the contradiction by preparing compensatory data in advance that can replace lost transmissions without requiring real-time communication during dropout events.
Solution Approach 2:
The system creates copies of control operations by generating multiple predicted future operations based on current state analysis. These predicted operations serve as substitutes for the original data stream during dropout events. The copying principle resolves the contradiction by creating redundant control sequences that can be activated when the original data transmission fails.
2Reliability
If machine learning modules predict future operations, then system functionality is maintained during interruptions, but computational complexity increases
Solution Approach 1:
The machine learning prediction system is segmented into multiple independent modules, each responsible for predicting a specific future operation. This modular architecture allows the system to maintain reliability through distributed prediction capabilities while managing computational complexity by dividing the overall prediction task into smaller, more manageable segments that can be processed independently.
3Measurement precision
If validity determination is performed on each message, then data quality is ensured, but processing time increases
Solution Approach 1:
Validity determination is performed as a preliminary action on each incoming message before it is processed further or used for predictions. This early validation ensures data quality is maintained while minimizing processing time by filtering out invalid data at the earliest possible stage, preventing wasted computation on corrupted messages and enabling the system to quickly identify when dropout events occur.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Apparatus and associated methods relate to mitigating data-stream dropout in a serial data-stream. A time-sequence of messages of the serial data-stream is received, each containing a data packet communicating an action. Validity of each of the time-sequence of messages received is determined. After receiving each valid message, a plurality of future actions is created based at least in part on the valid message received. The plurality of future actions corresponds to a plurality of future data packets of the time-sequence of messages. After receiving each valid message, the action communicated in the valid message received is performed. After receiving each invalid messages, a next one of the set of sequential future actions created is instead used in place of any action communicated in the data packet of the invalid message received.