Application of Fuzzy Rule based System for Highway Research Board Classification of Soils

TL;DR

This paper presents a fuzzy rule-based system for classifying soils in highway engineering, using fuzzy logic and decision trees to improve interpretability and match laboratory results.

Contribution

It introduces a novel fuzzy rule-based model for soil classification that integrates fuzzy logic with decision tree algorithms for better engineering assessment.

Findings

Fuzzy classification aligns well with laboratory test results.

The model effectively quantifies qualitative soil properties.

Implementation in MATLAB demonstrates practical applicability.

Abstract

Fuzzy rule-based model is a powerful tool for imitating the human way of thinking and solving uncertainty-related problems as it allows for understandable and interpretable rule bases. The objective of this paper is to study the applicability of fuzzy rule-based modelling to quantify soil classification for engineering purposes by qualitatively considering soil index properties. The classification system of the Highway Research Board is considered to illustrate a fuzzy rule-based model. The soil's index properties are fuzzified using triangular functions, and the fuzzy membership values are calculated. Fuzzy arithmetical operators are then applied to the membership values obtained for classification. Fuzzy decision tree classification algorithm is used to derive fuzzy if-then rules to quantify qualitative soil classification. The proposed system is implemented in MATLAB. The results obtained are checked and the implementation of the proposed model is measured against the outcomes of the laboratory tests.