Towards a parametric and context-realistic compaction training simulator

Sajad Mowlaei

Research output: ThesisEngD ThesisAcademic

60 Downloads (Pure)

Abstract

Virtual reality (VR) has been used for training in many industries and it already became a popular platform for education. VR provides a highly interactive environment that offers a unique opportunity to develop skillsets that are otherwise difficult, unsafe or costly to acquire in the real-world settings. VR-based training simulators have long been used in the construction sector for the training of equipment operators as well as workers. There are already advanced commercial simulators that support excavation, crane, and grading operations. There are also several simulators that are designed to sensitize general construction labor to safety rules and regulations. Trainees use these simulators to navigate in the VR environment and perform certain tasks that are designed to hone craftsmanship, dexterity, productivity, and safety skills.
Despite the growing popularity of training simulators, there are a few limitations that cast a shadow on the wide-spread applicability of them in the industry. (1) there are currently no training simulators for the asphalt paving operations. This is a major oversight because road construction constitutes a considerable portion of all civil engineering projects and hot mix asphalt is the dominant material in our roads. Besides, the proper compaction of asphalt is highly complex and requires very rigorous training and practice. Conventional training is very costly given the cost of equipment, material and space needed for training; (2) Existing simulators function based on pre-defined scenarios that are predominantly designed by developers with little to no affinity with paving operations. This reduces the applicability of simulators for a wide range of scenarios for which trainees need to be prepared; (3) Scenarios are commonly not realistic and fail to represent uncertainties and variabilities inherent in construction projects.
Based on the above problem, this design research intends to develop a framework for parametric and context-realistic compaction training simulators that can offer customizable and highly realistic training scenarios. To this end, a design research methodology was pursued wherein the stakeholder analysis was used to determine the functional requirements of the training simulator. Based on these requirements, an elaborate framework was developed to base a compaction training scenario on (a) actual data from real projects, and (b) inputs from the instructors that define the parameters of the training environment on the specific needs of target trainees. The framework was then implemented in a prototype system. In the developed simulator, the instructors can input the data from the actual construction site to represent the realistic cooling behaviour of the asphalt. They can also define the core parameters of the scene, such as the type of road, number of lanes, thickness, weather condition, number of required compaction passes, and the compaction temperature window. The developed prototype was tested in a number of cases to assess its usability, applicability and scalability.
The results of the validation suggest that the context-realistic and parametric training simulator has a great potential for progressing the training conventions in the road construction industry. The simulator is also found to be highly versatile, with applications that go beyond training and education. It is shown that the simulator can be used for project review (i.e., quality control), project planning, and Virtual Prototyping.
Original languageEnglish
Awarding Institution
  • University of Twente
Supervisors/Advisors
  • Dorée, Andries G., Supervisor
  • Vahdatikhaki, Faridaddin, Supervisor
  • Kolloffel, Bas Jan, Supervisor
Award date21 Mar 2023
Place of PublicationEnschede
Publisher
Publication statusPublished - 21 Mar 2023

Fingerprint

Dive into the research topics of 'Towards a parametric and context-realistic compaction training simulator'. Together they form a unique fingerprint.

Cite this