Constant False Alarm Rate (CFAR) Autotrend Evaluation Report
Abstract
A collaborative program was successfully conducted by Humaware and the Aviation and Missile Research, Development, and Engineering Center (AMRDEC) (managed by U.S. Army International Technology Center- Atlantic (USAITC-A)) between October 2010 and February 2011 to assess the performance of the Constant False Alarm Rate (CFAR) Autotrend dynamic alert detection technology as an augmentation to the Apache Modernized Signal Processor Unit (MSPU) fault detection technology. After an initial setup and optimization of the processing parameters, these results were achieved with a single set of parameters to control the CFAR and Autotrend processing, indicating a robustness that should result in a low overhead for data management. The evaluation data set contained events for all Condition Indicators (CIs) and for a large number of modules for the MSPU equipped AH-64D Apache helicopter. There was one set of CIs for the Auxiliary Power Unit (APU) that clearly indicated significant data quality issues, and these are being investigated by the Army. The principal findings of the assessment are that the CFAR Autotrend technology- (1) Exceeds the fixed threshold detection accuracy; (2) provides a significantly earlier detection than the fixed thresholds; (3) detects maintenance events (material faults) that are not identified by fixed thresholds; and (4) is ready for use as na engineering tool in its current form.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2011
- Accession Number
- ADA553586
Entities
People
- Daniel Wade
- Kenneth Pipe
- Steve Krick