Date of Graduation

Spring 2017


Master of Science in Project Management


Technology and Construction Management

Committee Chair

Amir H. Behzadan


The construction industry has one of the lowest productivity rates of all industries. To remedy this problem, project managers tend to increase personnel's workload (growing output), or assign more (often insufficiently trained) workers to certain tasks (reducing time). This, however, can expose personnel to work-related musculoskeletal disorders which if sustained over time, lead to health problems and financial loss. This Thesis presents a scientific methodology for collecting time-motion data via smartphone sensors, and analyzing the data for rigorous health and productivity assessment, thus creating new opportunities in research and development within the architecture, engineering, and construction (AEC) domain. In particular, first, a novel hypothesis is proposed for predicting features of a given body posture, followed by an equation for measuring trunk and shoulder flexions. Experimental results demonstrate that for eleven of the thirteen postures, calculated risk levels are identical to true values. Next, a machine learning-based methodology was designed and tested to calculate workers' productivity as well as ergonomic risks due to overexertion. Results show that calculated productivity values are in very close agreement with true values, and all calculated risk levels are identical to actual values. The presented data collection and analysis framework has a great potential to improve existing practices in construction and other domains by overcoming challenges associated with manual observations and direct measurement techniques.


ergonomics, productivity, wearable sensor, smartphone, machine learning, awkward posture, overexertion

Subject Categories

Business Administration, Management, and Operations


© Nipun Deb Nath

Open Access