Acquired Intelligence Inc.

Applications: Radio Interference AdvisorTM

A knowledge-based decision support system for radio inspectors in Canada.

The Information Problem

In most countries, the federal government assumes responsibility for management of the radio spectrum. In Canada, this responsibility includes responding to between 10,000 and 15,000 radio interference complaints annually. This responsibility is handled by 200 inspectors in 40 district offices.

The array of potential radio interference complaints that a radio inspector faces is enormous. The difficulty in solving these problems range from routine to exceptionally complex. Consequently, investigations can be very time consuming and costly for both the responsible federal department and the businesses and individuals affected by the interference.

Existing Practice

To assist with this responsibility radio inspectors in Canada have been using a spectrum management database system called SCOMS to record interference complaints and their solution. This database represents a wealth of anecdotal diagnostic information. However, this resource of information has not been as useful as it could have been in assisting inspectors with problematic investigations because there was no convenient way for them to access relevant investigations.

A New Approach

Glen Lockwood, Manager of New Technology Assessment in the Engineering Programs Branch of Industry Canada, felt that solution of these problematic investigations could be expedited through development of a system that could match an investigators current problematic situation to a database of historical, solved cases and present the most similar ones to the investigator for consideration. Given this, the Radio Interference AdvisorTM project was initiated.

The Radio Interference AdvisorTM

The Radio Interference AdvisorTM knowledge base, developed using Acquire®, has been organized into four major sections. Radio interference involving AM, FM, TV and RC (the commercial world of radio communications) make up these four sections. Within each of these sections the knowledge base has been further organized into 33 subsections according to the presenting symptom for the major section.

The knowledge base consists of 739 Objects organized into 498 Action Table rules (decision-table-like rules) to provide for inferencing. These 498 Action Table rules are roughly equivalent to 1,500 production rules.

Concurrent with the development of this knowledge base was the development of the Radio Interference Advisor case-based reasoning system and case management system that encompasses an embedded version of the inference engine and the knowledge base.

Given the extremely rare occurrence of certain symptoms and interference sources it was decided that the user should have access to both the knowledge-based systems conclusions/recommendations as well as the similar cases from the case-based reasoning system. In this way the user obtains guidance even if no case matches current circumstances. When matches are found the user is able to compare the advice from the knowledge base with that found in the similar cases and then take whatever action the user deems appropriate for their circumstance.

Conclusion

The utility of the Radio Interference Advisor is derived from three components: a comprehensive knowledge base covering radio interference complaints and their solution; a database of cases that provide the anecdotal reports on previous investigations; and a case-based reasoning system to bring these two components into a selection of similar cases for the consideration of a user with a problematic radio interference complaint. Coupling the Artificial Intelligence systems with the written reports from radio interference inspectors increases the applicability of the system and ensures that the system provides its solution in a form that is familiar and understandable by its users.
Further Reading
Applications: Radio Inspector's Assistant
Turn on the Radio: The RIA in AI Expert
AI Assists Radio Interference Inspectors: The RIA in Argus
AI Helps Radio Inspectors: The RIA in the ITU Handbook