Microburst Automatic Detection (MAD)
A patented system incorporates an algorithm that detects and quantifies microburst activity within a radar image field on a real-time basis.
University Corporation for Atmospheric Research, Boulder, Colorado
The portable, easily expandable Microburst Automatic Detection (MAD) system can dramatically improve aviation safety. Mimicking the reasoning of a human observer on the lookout for weather hazards, MAD assesses such hazards quickly while avoiding human weaknesses such as fatigue. Timely alerts to pilots and air-traffic controllers allow traffic rerouting to begin sooner to save lives and property. MAD's high success rate in confirmed detections and extremely low false-positive rate, and its utility with various radar and radar data formats, make the system versatile and applicable as a radar upgrade package.
MAD is a simple, user-friendly system with a small number of adjustable parameters that allow for easy modification of the algorithm to better process data for a specific location or prevailing conditions. Fuzzy logic data analysis techniques are used to detect and identify the size and location of microburst activity within a radar image on a real-time basis, presenting the user with likelihood images of such activity.
The particular benefits of the fuzzy logic approach mean that the invention can be readily augmented with additional inputs, and thus refined and improved as new inputs become available. For example, MAD is adapted to detect both terrain-induced wind shear in clear air, and severe and moderate turbulence as part of a larger weather warning system. Another augmentation would be to feed output from a second radar or another meteorological measurement device into an expanded MAD algorithm system.
In action, MAD accepts a series of low-level radar scans and converts these values to two-dimensional likelihood images for shear, storms, and clutter, each image defining or distinguishing some characteristic of a microburst. A radar scan is transformed into a likelihood image through likelihood mapping, which uses input fields that include, but are not limited to, radial velocity, reflectivity, wind-shear estimates, and clutter maps. The combined likelihood image is then processed with pattern-matching techniques to produce a final smoothed likelihood image. The set of point locations of interest that are above a predetermined threshold in the final likelihood images are built into connected regions, whose boundaries define a microburst footprint. Such footprints are represented as polygons and can be overlaid on a polar or Cartesian coordinate map of the airport region.
Developers of MAD have also noted significant parallels between decision-making based upon a radar image and medical imaging technologies, concluding that there may be strong potential for application of the MAD algorithm system to the computer-aided diagnosis of breast cancers. A system would be based on fuzzy logic data-weighting analysis to process CAT scan, x-ray, or MRI scan image data and other medical data automatically to quickly alert a radiologist or physician to the location of malignancies. In particular, higher accuracy and lower false-positive/negative identifications could improve diagnosis and save lives.
The Microburst Automatic Detection (MAD ) algorithm flowchart.
The MAD system was developed by Dave Albo and Dr. Kent Goodrich of the Research Applications Program at the University Corporation for Atmospheric Research, sponsored by the Federal Aviation Administration and the National Science Foundation. Inquiries concerning rights for the commercial use of this invention should be addressed to Technology Commercialization Program, University Corporation for Atmospheric Research Foundation, PO Box 3000, Boulder, CO 80307; (303) 497-8588; fax: (303) 497-8561; E-mail: IPMP_Frontdesk@qgate.ucar.edu; URL: http://www.ucar.edu.