Models For Predicting RWIS Sensor Misalignments and Their Causes - FY07 NATSRL

Principal Investigator(s):

Donald Crouch, Former Professor, UMD-Electrical Engineering

Co-Investigators:

Project summary:

The objective of this project was to develop a pilot software tool that allowed a Road/Weather Information System (RWIS) site engineer to automatically assess a group of RWIS sites and determine when malfunctions occur. Specifically, this project developed an interface that automated the process of assessing the set of weather sensors being maintained by a single technician. In a recent survey of RWIS maintenance personnel, a full 50% of the respondents indicated they were not at all satisfied with RWIS equipment reliability and data accuracy. Some 40% indicated a need for additional training and resources for maintenance; at least half of the lead technicians indicated they were not at all satisfied with the support provided. The goals of the initial phase of the project were: 1) to capture the existing process of a Minnesota Department of Transportation technician; and 2) to demonstrate that the software can recognize malfunctions that are identifiable by that technician. One motivation for this work was that automating the analysis process would make it possible to export that process to technicians responsible for similar sensors in other parts of the state and even make it possible for technicians who only work part-time on the sensors (and therefore may not be as expert in the process of assessing the sensors) to effectively assess the sensors for malfunctions. This required capturing the expertise of the existing technicians to allow this knowledge to be saved and retained. A second motivation for the initial phase was to make the process of assessing a group of sensors easier for each technician, and to make it possible for the notification mechanism to be incorporated into the transportation agency's central mechanism. The final report provides an overview of the software tool developed to support detection of sensor malfunctions.

Sponsor(s):

Project details:

Reports or Products: