Software for the detection of anomalies, incidents, and changes in sensor signal data
- Learning expert system for detection of anomalies, incidents, and changes in signal data
- Comprehensive set of criteria with different complexities for identification of anomalies
- Quality assessment for machine learning training and validation data with full traceability
- Embeddable into Simulation Data Management or test management systems
- Flexible and automated reporting and data repair mechanisms
Expectator is a software suite for the automated plausibilisation and detection of errors in numeric data. It is designed specifically for the identification of anomalies and incidents within sensor signals and time series data of arbitrary types.
Anomalies are, by definition, unexpected, and we do not believe in a one-size-fits-all algorithm for incident and anomaly detection. Instead, Expectator follows the philosophy of defining simple and pragmatic criteria for the description of the expected regular characteristics and shape of sensor signals.
It combines these numerous criteria to create powerful committees resulting fewer wrong classifications (false positives), retaining good detection rates. Several methods with different complexities and functionalities are available.
Expectator is a tool for the quick and pragmatic development, formulation, definition, calibration, administration, evaluation, and adaptation of so-called expectations to the content of arbitrary sensor signals. For this, we distinguish between the generic tool for editing and administrating the expectations (offline) and specialized tools for executing and evaluating the expectations in the embedded target environment (online). The latter is usually customized for the corresponding application domain and environment.
Expectator provides several features, including:
- Definition, calibration, and expectations development of signals using various methods
- Determining simple statistical attributes of single and combined signals
- Finding signal characteristics in the frequency domain for transient and periodic signals
- Determining if signals cross defined intervals or lie within certain bands
- Finding the distance between two signals or from one signal to a reference signal
- Using Self Organizing Maps for the identification of anomalies
- Checking the difference between real measurements and predictive algorithms
- Checking signal patterns with neural networks and deep learning methods
- Customer and application-specific adoption and integration of the evaluation for arbitrary target platforms
- Expandable with arbitrary expectation methods for data plausibilisation and calibration in expert systems
Expectator is a MATLAB toolbox based on ANDATA's Signal Structure Toolbox. All functions of Expectator can be controlled via a command line interface. This allows all operations to be combined into application-specific scripts or apps and integrated into specific work environments. The product is useful for users dealing with huge amounts of signal data from test and/or simulations, and for experts who are automating processes and workflows with signal data, such as test automation systems or simulation data management.
- On-site assistance
- System integration
- Data Analysis Tools
- Data Acquisition or Import
- Data Analysis and Statistics
- Digital Signal Processing
- Process Control and Monitoring
- System Modeling and Simulation
- Aerospace and Defense
- Industrial Automation and Machinery
- Rail, Ships, and Other Transportation
- Utilities and Energy
Related Connections Views: Data Analysis Tools, Data Acquisition or Import, Data Analysis and Statistics, Digital Signal Processing, Process Control and Monitoring, System Modeling and Simulation, Aerospace and Defense, Automotive, Industrial Automation and Machinery, Rail, Ships, and Other Transportation, Utilities and Energy