Study of retrofitted system for Intelligent Compaction Analyzer - a Machine Learning approach for Quality Control of Asphalt Pavement during Construction

Aug 5, 2026

Asphalt pavements play a critical role in transportation infrastructure, yet construction practices can yield poor quality and weaken performance. This thesis presents the Retrofit Intelligent Compaction Analyzer (RICA), a real-time system that estimates compaction density during asphalt placement. RICA applies machine learning to predict density from received vibratory patterns at multiple compaction levels, using the roller’s spatial location and vibration signal analysis. Field data from an actual construction site were used, and RICA estimates were validated against roadway core measurements, confirming reliable density estimation and improved quality assurance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages127

Reading Time5 h 20 min

AccessFree to download

Introduction

Problem Statement

Literature Review

Scope and Novelty

Organization

Data Collection

Instrumentation

Retrofit system

Structure of raw data

Collection of ground truth

Collection of calibration data

Feature Engineering

Data Imputation

Derivation of power attributes from vibration data

Formulation of rolling pattern attributes

Normalization

Removal of outliers

Feature Extraction

Data Visualization

Density Prediction

Learning mechanism

University of Nevada, Reno Study of retrofitted system for Intelligent Compaction Analyzer, a Machine Learning approach for Quality Control of Asphalt Pavement during Construction A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science and Engineering by Shankar Poudel Dr. Sesh Commuri and Dr. George Bebis-Thesis Advisors August, 2023 Copyright © 2023 Shankar Poudel: All rights reserved. THE GRADUATE SCHOOL We recommend that the thesis prepared under our supervision by Shankar Poudel entitled Study of retrofitted system for Intelligent Compaction Analyzer, a Machine Learning approach for Quality Control of Asphalt be accepted in partial fulfillment of the requirements for the degree of Master of Science George Bebis, Ph.D Advisor Sesh Commuri, Ph.D. Co-advisor Alireza Tavakkoli. Ph. D. Committee Member Jeongwon Park, Ph. D. Graduate School Representative Markus Kemmelmeier, Ph.D., Dean Graduate School August 2023 i ABSTRACT Asphalt pavements play a vital role in transportation infrastructure, but their performance can suffer due to subpar quality resulting from improper construction practices. To tackle this issue, we introduce the Retrofit Intelligent Compaction Analyzer (RICA), a real-time compaction density estimation system for asphalt pavements during construction. RICA utilizes machine learning principles and machine learning to predict compaction density based on received vibratory patterns at different compaction levels. By leveraging the roller’s spatial location and analyzing vibration patterns, RICA delivers density estimates. In this study, we gathered data from actual construction sites, implementing RICA on a Caterpillar CB-10 Rotary dialed dual drum vibratory compactor. The density estimates from RICA were validated against densities measured from roadway cores extracted randomly on the compacted pavement. Our findings affirm the efficacy of RICA in providing reliable density estimates for asphalt pavements. The ability of RICA to provide real-time, nondestructive compaction information to the roller operator establishes its value as a quality control tool during asphalt pavement construction. By ensuring proper compaction, RICA contributes to the construction of durable, high-quality roads while reducing the financial and environmental costs associated with construction and maintenance. The validation of RICA’s estimates with percent within limits (PWL) calculations based on roadway cores further attests to its effectiveness as a Quality Assurance tool. Keywords: Intelligent asphalt compaction analyzer, density estimation, machine learning, compaction quality, nondestructive testing, quality assurance ii ACKNOWLEDGEMENTS I would like to express my deepest gratitude to my advisors, Dr. George Bebisand Dr. Sesh Commuri, for their invaluable patience and constructive feedback throughout this research journey. Their guidance and expertise have been instrumental in shaping this thesis. I am also grateful to my defense committee for their valuable insights and contributions to this work. Their knowledge and expertise have been crucial in refining the ideas presented in this thesis. I extend my appreciation to my friends, lab mates, librarians, and research assistants at the university. Their support, encouragement, and inspiration have been invaluable during various stages of this work. Special thanks go to George Reed Inc. for their generous support in facilitating this research. Their provision of sites and the necessary equipment, including the rollers for field data collection, has been instrumental in the success of this project. I am also thankful to Dr. Garrett Winkelmaier for his feedback sessions, moral support, and encouragement throughout the research process. I would not have been able to complete this endeavor without his help and assistance. Lastly, I would like to acknow

FAQs

What problem does the thesis address in asphalt pavement construction quality?

Asphalt pavement performance can suffer when compaction quality is inadequate due to improper construction practices. The thesis targets reliable, real-time density information during construction to support quality control.

What is RICA and how does it estimate compaction density?

RICA (Retrofit Intelligent Compaction Analyzer) is a real-time system that uses machine learning to predict compaction density from vibratory patterns at different compaction levels. It also leverages the roller’s spatial location while analyzing vibration signals.

How was RICA validated in the study?

Data were gathered from real construction sites by installing RICA on a Caterpillar CB-10 dual-drum vibratory compactor. Density estimates from RICA were validated against densities measured from randomly extracted roadway cores, including percent within limits (PWL) comparisons.

Aug 5, 2026

Upload by Theodore