Diagnosing PARC’s Refrigerator Benchmark with Data-Driven Methods
Details
Denver, CA USA. Date of Talk: 2016-10-04
Speakers
Event
Diagnosing PARC’s Refrigerator Benchmark with Data-Driven Methods
Diagnosing real-world systems is a challenging task due to, e.g., complexity, measurement error and lack of control in experimentation. Designing controlled experiments, diagnostic algorithms, and building benchmarks to validate the algorithms pave the way to address the challenge. More than 1800 hours of power and temperature data from three identical household refrigerators, modified for the purpose of benchmark creation, were recorded. Controlled failure injection of non-destructive faults has been performed over multiple 24-hour scenarios. The thermal and electric data in the benchmark has been analyzed manually and with four data-driven machine-learning-based algorithms. The performance of each data-driven, machine-learning diagnostic algorithm has been characterized in terms of diagnostic metrics such as false positives, false negatives, and isolation time.
Additional information
Focus Areas
Our work is centered around a series of Focus Areas that we believe are the future of science and technology.
Licensing & Commercialization Opportunities
We’re continually developing new technologies, many of which are available for Commercialization.
News
PARC scientists and staffers are active members and contributors to the science and technology communities.