Qualified Anomaly Detection in Serial Data Streams

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for analyzing serial data streams struggle to accurately detect and trigger on specific analog anomalies due to noise and complexity, particularly in debugging physical events within hierarchical serial data streams, where analog anomalies can lead to misinterpretation of bits.

Innovation Solution

A method that detects a start pattern, such as a packet header, to generate an enable signal and a stop event, like a packet trailer, to produce a qualification signal, allowing a trigger circuit to capture data surrounding specified anomalies within a defined portion of the serial data stream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If trigger circuits are made complex to enable detection of any analog anomaly, then detection capability is improved, but device complexity increases

Engineering Contradiction:

Improveanomaly detection capabilityVSAvoidtrigger circuit complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the serial data stream into hierarchical levels (physical layer, data link layer, etc.) and defines qualification criteria at each level. The trigger circuit is divided into functional modules: a qualification signal generator that processes hierarchical data to define a qualified area, and a trigger circuit that processes analog anomalies only within that qualified area. This segmentation reduces overall circuit complexity while maintaining comprehensive detection capability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary action by first generating a qualification signal before triggering the anomaly detection. The qualification signal is generated by processing serial data through hierarchical levels to identify a qualified area, and only after this preliminary qualification does the trigger circuit become active for detecting analog anomalies. This preliminary filtering action reduces the burden on the trigger circuit.

Inventive Principle:

Principle #10Preliminary action

2Difficulty of detecting and measuring

If analog anomalies are detected in noisy serial data streams, then anomaly detection is enabled, but measurement precision deteriorates due to misinterpretation

Engineering Contradiction:

Improveanomaly detection capabilityVSAvoidbit interpretation accuracy

Core Design Contradiction:

Difficulty of detecting and measuringVSMeasurement precision

Solution Approach 1:

The patent performs preliminary action by first establishing a qualification signal that defines a qualified area before attempting to detect analog anomalies. The qualification signal is generated by processing serial data through hierarchical levels (physical layer, data link layer protocols) to identify valid data segments. Only within this pre-qualified area does the trigger circuit search for analog anomalies, ensuring that decoding occurs only on validated data portions, thereby preventing misinterpretation of noisy bits.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary mechanism - the qualification signal - that mediates between the noisy serial data stream and the anomaly detection process. The qualification signal acts as a filter and validator, ensuring that only data within the qualified area (defined by start and stop events at hierarchical levels) is subjected to anomaly detection. This intermediary prevents direct interpretation of noisy bits outside the qualified area.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Adaptability or versatility

If the entire serial data stream is analyzed for anomalies, then detection coverage is improved, but processing time increases

Engineering Contradiction:

Improvedetection coverageVSAvoidprocessing time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the serial data stream into hierarchical levels (physical layer, data link layer, etc.) and uses qualification signals to identify specific segments of interest. Instead of analyzing the entire stream, the system processes data at hierarchical levels to define qualified areas bounded by start and stop events. This segmentation enables selective anomaly detection only within relevant segments, reducing processing time while maintaining comprehensive coverage across different protocol layers.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary action by first identifying the qualified area through hierarchical data processing before performing anomaly detection. The qualification signal is generated in advance by processing serial data through physical layer and data link layer protocols to identify valid segments. This preliminary identification of relevant areas reduces the amount of data that requires subsequent anomaly processing, thereby reducing overall processing time while maintaining detection coverage.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS7480839B2Qualified anomaly detection

Publication Date: 2009.01.20 TEKTRONIX INC

AI summary

A circuit and method of qualified anomaly detection provides detection and triggering on specific analog anomalies and/or digital data within a qualified area of a serial data stream. A start pattern within the serial data stream, such as a packet header, is detected to generate an enable signal. A stop event, such as a packet trailer, a specified digital event, a time interval or the like, is identified to generate a disable signal. The enable and disable signals are combined to produce a qualification signal that allows a trigger circuit to trigger on a specified anomaly within the portion of the serial data stream defined by the qualification signal.