The laboratory course is a co-operation of multiple departments. Each student performs two MIT and five (IIT or AUT) measurements depending on the chosen specialization.

 

This website contains information about the two MIT measurements and materials, all other official information can be found on the official website of the course.

 

Topic of the lab

The goal of the laboratory work task is to learn some practice in knowledge based expert system development, more specifically in rule based system design and implementation. You should become an „expert” for a specified domain in your choice, and develop a simple system that provides solution for questions for non-expert users.

 

Task of the laboratory

Create a rule based system that provide advices in a specific topic in the domain of traffic rules, e.g. how to pass a crossing, where it is admitted to park, and so on. The knowledge base should contain at least 12 rules and your system should be able to question the user about the details of the problem he is facing and support him an advice at the end of the discussion.

 

Submission of the work

All documents should be submitted electronically to the HF Portal of the department (https://hf.mit.bme.hu).

You should upload all of your documents not later than 18^th of November 2018:

• documentation of your solution (idea, brief description of the knowledge base, sample dialogue, evaluation of the results; 1-2 page document is exected,
• source code of your program

All answers must be your own work and you are not allowed to work together with other students in completing any part of this assignment. Nor are you allowed to discuss the answers with other students. Your answers will be subjected to plagiarism-detection software.

Development instructions

You can use any expert system tool or programming language to solve the task, but we suggest you to use the CLIPS framework that was introduced on the lectures.

CLIPS framework and all the documentation can be found at: http://clipsrules.sourceforge.net/

We suggest you to analyse and use the example programs that are provided, especially auto.clp for feedforward reasoning or animal.clp for backward chaining.

It is also your choice what domain you choose, we provide you some example at the beginning of the laboratory event.