Automation in paving operations, too little or too much? - Effectiveness and usability assessment of implementing different degrees of automation in paving operator support systems
Denis Makarov is a PhD student in the department Construction Management and Engineering. (Co)Promotors are prof.dr.ir. A.G. Doree, dr. S.R. Miller and dr.ir. F. Vahdatikhaki from the faculty Engineering Technology.
In the modern world, the development of societies and their road infrastructure are inseparable. With the rise of global challenges, such as urbanization, climate change, and economic growth, the importance of roads to facilitate the movement of goods and people cannot be overstated. To ensure that road networks play their critical role in promoting social and economic progress, it is vital to maintain high standards in road construction. However, this is becoming increasingly challenging given the rapid pace of urbanization, which has resulted in higher traffic loads, faster degradation, and greater maintenance costs. Hence, it is imperative to construct roads that can endure for their intended lifespan.
In numerous countries, Hot Mixed Asphalt (HMA) is the favored and primary substance used to construct roads. To ensure top-notch HMA roads, proper compaction is essential. This entails considering both the degree of compaction and the temperature of the asphalt simultaneously. Neglecting these factors could lead to asphalt that is either under or over-compacted.
The compaction process of HMA is equipment-intensive, and the interdependency between the machines involved makes process control difficult and introduces process variability.
Operators and workers play a significant role in this process, but conventional training and reliance on intuition or tacit knowledge can intensify variability and affect productivity, quality, and safety.
Addressing these challenges requires real-time systems to guide operators and improve performance. This can be seen in other industries that have successfully used digital technologies and automation to improve quality and sustainability. Automation can improve productivity and reduce the burden on the workforce. This makes the industry more appealing to younger, technologically-savvy generations. This transformation can also change the industry's image from being a low-skill and conventional sector to an advanced working environment, encouraging more people to join the industry.
In the Netherlands, public clients have introduced new contract forms and extended warranty periods to improve road quality and durability by reducing process variability and making explicit operational practices. While contractors are adopting digital technologies, such as Global Positioning Systems (GPS) and infrared thermography to improve construction quality, their use is still limited.
The aging workforce in the construction industry and a lack of young entrants, combined with increasingly complex tasks, underscore the need for different approaches to develop and use Operator Support Systems (OSSs). OSSs use advanced technologies, such as the Internet-of-Things (IoT) and Machine-to-Machine (M2M) communication to collect and share data and provide process-related information to equipment operators in real-time.
However, current OSSs are mostly descriptive (collecting and presenting data). There is little development in semi-prescriptive and prescriptive OSSs, systems that partially or fully process collected data and guide operator(s). This is due to low adoption rates and the lack of specific metrics to measure quality in paving operations.
While the industry aspires to move towards more autonomous operations, it is unclear how to make the transition both technically and receptive for users. This uncertainty raises questions about the success of the industry's automation ‘gold rush’ mindset and its ambitions to move to fully autonomous equipment. Given the current level of autonomy and other barriers, this thesis deals with the pressing question – how to migrate from descriptive to more prescriptive OSSs?
While paving OSSs have been developed and implemented in construction practices, there is a notable absence of knowledge as to how to advance beyond the low level of automation (Level 1). Any plans to create and envision higher levels of automation must consider two critical developmental factors: (1) how the transition will enhance the quality of the operations (i.e., objective verification of value); and, (2) how the transition will align with the needs of the end-users (i.e., user reception).
Accordingly, this research aimed to address the following research question:
“How different levels of automation in paving OSSs can be implemented and what implications they will have on the quality of the paving operation and on the perception of the operators?”
To answer this question, the thesis suggests different methods to: (1) construct the necessary hardware framework capable of providing varying levels of automation to operators engaged in asphalt paving and compaction in real-time; (2) translate raw sensory data from descriptive paving OSS to semi-prescriptive support; (3) transform unprocessed sensory information extracted from the descriptive paving OSS into a prescriptive format; and, (4) evaluate the readiness of road construction machines operators to transition towards more advanced levels of automation.
These proposed methods underwent iterative development, execution, and examination, with the assistance of specialists from the asphalt paving industry.
The first method provides an overview of the essential hardware components of the paving OSS setup that facilitates different levels of operator support and process automation. The prototype system developed in this phase of the research addresses the primary research challenge: i.e., how to create a new type of paving OSS that considers the actors, activities, and products involved and leverages real-time data on process conditions?
Next, the method for the descriptive OSS was developed. This part of the research illustrates the application and benefits of this level of automation using a case study conducted on actual construction sites.
Subsequently, a semi-prescriptive OSS was developed that supports operators by providing them with compaction priority maps. Semi-prescriptive OSS translates the raw data gathered by the descriptive system into actionable guidance. This requires less interpretation by the machine operators. The priority mapping method designed for this purpose was evaluated using a novel metric known as the Effective Compaction Rate (ECR). This allows an objective and quantitative assessment of the quality of the compaction operation.
The findings of this method led to the development and testing in the next step of a novel compaction trajectory planning method that can be used to implement prescriptive support in paving OSSs.
The three methods for descriptive, semi-prescriptive, and prescriptive support provide a comprehensive understanding of the possible trajectory for the evolution of paving OSSs.
The last research step took a critical look at the usability aspect of the different automation levels of OSS. To overcome the issues associated with traditional approaches that use physical prototypes, a virtual prototype (VP) platform was developed and used as a medium for usability testing.
The research presented establishes a comprehensive framework for technology transition in the field of paving OSSs. It provides, not only technical insights into various ways OSSs can support operators, but also addresses technology adoption and reception issues.
Previously, construction equipment OSS developments lacked clear classification in terms of operator support levels. This research, for the first time, adapts the concept of levels of automation from the aviation and automobile industries to construction equipment OSS. It offers a guiding framework to analyze the current state and strategize future advancements.
The research introduces a distinct assessment metric to evaluate paving OSSs objectively. Unlike previous methods that relied on indirect or low-resolution measurements, this novel metric considers operation homogeneity and compaction quality that align with optimal compaction strategies. The proposed metric has the potential to become the gold standard in paving OSS development and standardization, as it allows high-resolution measurement without additional equipment deployment.
The research dives into the discourse on levels of automation and technology adoption. It sheds light on the trade-off between technological intervention and acceptance. By adopting a socio-technical perspective, the study introduces a user-in-loop design and development methodology for paving OSSs. The innovative use of virtual prototyping as a safe technology assessment tool supports this approach. The research aims to sensitize researchers and practitioners to the multi-dimensionality and complexity involved in construction equipment OSS development.
Overall, this research contributes a robust framework, a specific assessment metric, and a user-in-loop methodology to advance the field of paving OSSs, addressing both technical and socio-technical aspects.
