Electronic productivity performance monitoring of construction workers

The Construction Industry (CI) is failing to boost its productivity. Moreover, 56% of the industry employment is craft workforce. This statistic means that increasing craft workforce productivity introduces positive impacts on overall construction productivity. The topic is relevant for many constituents from Government to the Companies level. The lack of craft workforce available in developed countries and the CI increasing requirements towards higher productivity and efficiency imply these agents' actions. The research frames specifically the craft workforce performance as the problem to be solved in the CI productivity scenario. Many authors have worked on the construction productivity theme (e. g. assessment levels, processes analysis, impact factors). Also, human activities recognition/classification applying sensing technology and automated processing have being experienced nowadays. This work combines both contributions to produce the outcomes that will provide the opportunity to boost on-site productivity. Higher competitiveness and efficiency are goals for the construction industry. Construction 4. 0 is a megatrend and a driver towards those objectives. There is a gap between the relevance of the craft workforce for overall productivity and the investment in actions to measure and improve that productivity. In a 4. 0 context, this gap tends to increase as low technology is invested in craft workers. This work highlights the relevance of the craft workforce on the overall industry and forecasts how a 4. 0 vision targeted for these agents can decrease the gap mentioned above. The thesis main aim is to develop a framework to measure craft workforce productivity. Being the objectives: Conceptualise an approach to assess and modelling craft workforce productivity; Provide in-depth knowledge to enable GDPR (General Data Protection Regulation) compliance; Conceptualise the Electronic Performance Monitoring (EPM) data flow, connecting system/methodologies; Testing the EPM application and validation of the proposed framework. The research is undertaken on an exploratory basis; this work "mix and match" methods to better achieve the particularities of this construction management research. This work combines new approaches for the CI to achieve 4. 0 era using a well-known methodology for motion study (that had little experimentation in this sector) and innovative processes for data collection (wearables sensing technology) and processing (machine-learning and multivariate statistical analysis to classification). Ten construction activities were simulated in laboratory facilities performed by six volunteers using wearables to collect the acceleration (two wrists and one leg) and applying machine-learning plus multivariate statistical analysis for post-processing. In addition, data were collected on-site (through work sampling) to measure and model two production processes typical of civil construction. Craft workers also carried out wearable electronic devices to collect their reactions to EPM through interviews. A survey containing twenty-five questions was proposed for collecting managers-level awareness about EPM at the workplace. Experts semi-structured interviews from three different countries were used to capture the conceptualisations and challenges regarding Worker 4. 0 in the CI. The primary outcome is a framework to model work processes and measure productivity. Craft performance modelling can bring awareness about the on-site tasks mechanization processes fostering productivity improvements. The research provides in-depth knowledge to enable GDPR compliance based on 18 packages of agreements. Also, conceptualise an EPM data flow, connecting system/methodologies. Based on the survey, an expressive level of 56% EPM acceptance was determined, and surprisingly 69% of the respondents first knew about EPM through the survey. An approach to assess and modelling craft workforce productivity was conceptualised as Worker 4. 0 Motion Productivity (Contenting nine processes: Free-hand performing, Auxiliary tools, Manual tools El ctric/Electronic tools, Machines operation, Robotic automation, Do not operating value, Walking, Carrying); Worker 4. 0 Flow to boost Productivity (Efficiency flow connecting the nine processes and accounting for the amount of performance on each; plus a Mechanization index level calculated based on the process modelling); Worker 4. 0 Mapping chart (Two-hand methodology to be fulfilled based on an EPM system).  The results applying EPM in the laboratory circuit for three IMUs (both wrists and one at the dominant leg) achieve accuracy between 92-96% for Machine-learning processing and 47-76% for the Multivariate statistical analysis approach. And, for one IMU (only wrist-dominant data), 32-76% accuracy for Multivariate statistical analysis approach. On-site experiments collecting data from three activities were performed, resulting in the modelling of the processes and fulfilling the Worker 4. 0 Flow to boost productivity, achieving: Water supply system (to consumers' taps) a Mechanization index of 42. 20% considered High in the scale proposed; Masonry (residential building with expanded clay blocks) a Mechanization index of 15. 69% considered Low in the scale proposed; Masonry (commercial building with expanded clay blocks and concrete blocks) a Mechanization index of 13. 64% considered Low in the scale proposed. Due to the restrictions imposed by the world pandemic of COVID-19, it was impossible to carry out further experiments. Such restrictions limited the development of the work. In the electronic monitoring field, only ten activities were catalogued among four of the nine idealised processes. It was also not possible to process data collected on-site electronically. To meet the need for process modelling, it was only possible to develop analyses at three different construction sites using a work sampling methodology. As a result, the amount of data/results was only limited to masonry and piping activities. It was impossible to perform any tests to validate the Worker 4. 0 Mapping chart concept, nor was it possible to supply any KPIs identified with collected data. The contribution to knowledge for theoretical implication includes the conceptualisation of a framework craft-workforce-centred focusing on the modelling of construction activities. The practical use of a quantitative Mechanization index can lead to a benchmarking of activities prone to delivering the highest levels of outcomes and increase CI productivity. The future direction of this work should focus on expanding on-site data collection. It should also seek to establish algorithms for automating data processing.