Citra Islami, Rizky (2026) MONITORING PRODUKTIVITAS PEKERJA KONSTRUKSI SECARA OTOMATIS BERBASIS DEEP LEARNING. Doctoral thesis, UNDIP.

	Text Hal. Pengesahan.pdf Download (1MB)
	Text BAB 1.pdf Restricted to Repository staff only Download (57kB)
	Text BAB 2.pdf Restricted to Repository staff only Download (4MB)
	Text BAB 3.pdf Restricted to Repository staff only Download (2MB)
	Text BAB 4.pdf Restricted to Repository staff only Download (960kB)
	Text BAB 5.pdf Restricted to Repository staff only Download (1MB)
	Text BAB 6.pdf Restricted to Repository staff only Download (87kB)
	Text DAFTAR PUSTAKA.pdf Restricted to Repository staff only Download (211kB)

Abstract

The construction industri is a labor-intensive industri, but construction worker
productivity remains low, and workers tend to do repetitive tasks that makes time�wasting. Monitoring of construction worker performance is still done manually, which is
time-consuming and carries the risk of recording errors. Performance monitoring with the
integration of technology and artificial intelligence (AI) that has been developed mostly
focuses on monitoring discipline and calculating tool and material usage. However, the
urgency in the field is to increase construction worker productivity to optimize projects
performance. This gap indicates the need to develop a more accurate and real-time
construction worker productivity evaluation method using AI. The purpose of this
research is to develop a deep learning-based construction worker productivity monitoring
model. The analysis stage is carried out in three stages, each using methods including, (1)
evaluating the implementation of AI for monitoring construction worker productivity
using the AI implementation readiness index by distributing questionnaires whose index
results are depicted using a spider web, (2) developing an automatic deep learning-based
construction worker monitoring model based on the You Only Look Once (YOLO)
algorithm that is run on the Roboflow and Google Colabs platforms, and (3) predicting
the construction worker productivity cycle time based on the model that has been built.
The case study used was a wall masonry project, which presents the potential for a wide
variety of movements that can be fully observed. The resulting model achieved an F1
score of 93.45%, indicating a reasonably good dataset recognition. Meanwhile, the
predicted productivity of construction workers was 122,3 seconds for one wall masonry
work cycle, which includes measuring, mixing mortar, and laying bricks. The findings of
this model are expected to contribute to assisting service providers in monitoring
performance to improve construction worker productivity.
Keywords: productivity, construction workers, automated technology, monitoring, deep
learning

Actions (login required)

View Item