Recognizing Sitting Activities of Excavator Operators - Using Multi-Sensor Data Fusion with Machine Learning and Deep Learning Algorithms

Aug 6, 2026

Upload by Finn

Recognizing excavator operators’ sitting activities is vital for strengthening health, safety, and productivity while also clarifying how operators behave and interact with construction equipment. Prior work is limited for this specific sitting-activity scenario, especially under real-site constraints. This study proposes a recognition method that fuses multi-sensor data and applies machine learning and deep learning models. A system combining interface pressure sensor arrays and inertial measurement units is deployed on a construction site, achieving strong accuracy for static postures and compound actions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages47

Reading Time1 h 58 min

AccessFree to download

Abstract

Introduction

Activity recognition background

Proposed multi-sensor learning approach

1 Recognizing Sitting Activities of Excavator Operators Using Multi-Sensor Data 2 Fusion with Machine Learning and Deep Learning Algorithms 3 Jue Lia, *, Gaotong Chena, Maxwell Fordjour Antwi-Afarib 4 a School of Economics and Management, China University of Geosciences, Wuhan, Hubei, China 5 b Department of Civil Engineering, College of Engineering and Physical Sciences, Aston University, Birmingham, 6 UK 7 * Corresponding author at: School of Economics and Management, China University of Geosciences, Wuhan, Hubei, 8 China. E-mail address: [jueli0824@gmail.com](jueli0824@gmail.com) 9 Abstract 10 Recognizing excavator operators’ sitting activities is crucial for improving their health, safety, and productivity. 11 Moreover, it provides essential information for comprehending operators’ behavior patterns and their interaction with 12 construction equipment. However, limited research has been conducted on recognizing excavator operators’ sitting 13 activities. This paper presents a method for recognizing excavator operators’ sitting activities by leveraging multi- 14 sensor data and employing machine learning and deep learning algorithms. A multi-sensor system integrating 15 interface pressure sensor arrays and inertial measurement units was developed to capture excavator operators’ sitting 16 activity information at a real construction site. Results suggest that the gated recurrent unit achieved outstanding 17 performance, with 98.50% accuracy for static sitting postures and 94.25% accuracy for compound sitting actions. 18 Moreover, several multi-sensor combination schemes were proposed to strike a balance between practicability and 19 recognition accuracy. These findings demonstrate the feasibility and potential of the proposed approach for 20 recognizing operators’ sitting activities on construction sites. 21 Keywords: Excavator operator; Sitting activity recognition; Multi-sensor fusion; Machine learning; Deep learning; 22 Interface pressure 23 1. Introduction 24 Construction equipment operation plays a vital role on construction sites, relying heavily on the expertise and 25 efficiency of construction equipment operators. These operators bear the responsibility of operating a wide range of 26 machinery, such as excavators, cranes, bulldozers, and forklifts, which are indispensable for ensuring the safe and 27 effective completion of construction projects. Throughout the operation of construction equipment, the activities 28 performed by operators dominate the operational process and affect various aspects of construction task performance 29 [1] . These activities encompass a diverse range of body postures and compound movements carried out by the 30 operator while seated, playing a crucial role in their daily tasks [2-5] . Operators often spend prolonged durations 31 operating construction equipment in harsh environments and are often subjected to conditions that pose significant 32 construction risks [2-5] . Notably, operators’ sitting activities have been demonstrated to possess profound 33 implications for construction performance, including operational safety [6,7], workers’ health [2,3,8,9], and 34 production efficiency [10] . Moreover, human body activities serve as fundamental elements reflecting human 35 behaviors during interactions with the environment [11] . With the growing presence of machinery operations with 36 varying levels of automation at modern construction sites [12], continuous monitoring of operators’ sitting activities 37 has become increasingly necessary. Such monitoring provides essential information for modeling and analyzing 38 operators’ daily operational performance and behavioral patterns, thereby offering insights for understanding human- 39 machine interactions context at the construction sites. Therefore, given the strong relationship between sitting 40 activities and construction equipment operation performance, it b

FAQs

Why is sitting activity recognition for excavator operators important?

It improves health, safety, and productivity, and provides information to understand operators’ behavior patterns and interaction with construction equipment.

What sensor system is used to recognize sitting activities in this study?

It integrates interface pressure sensor arrays and inertial measurement units to capture sitting-activity information at a real construction site.

Which modeling approach shows the best reported performance?

The gated recurrent unit achieves outstanding performance, with 98.50% accuracy for static sitting postures and 94.25% accuracy for compound sitting actions.

Aug 6, 2026

Upload by Finn