| Robert Amor's Publications in 2022 |
|---|
 |
Feng, Z., González, V.A., Mutch, C., Amor, R., Cabrera-Guerrero, G. (2022) Exploring Spiral Narratives with Immediate Feedback in Immersive Virtual Reality Serious Games for Earthquake Emergency Training, Multimedia Tools and Applications, https://doi.org/10.1007/s11042-022-13306-z. |
Abstract: Various attempts and approaches have been made to teach individuals about the knowledge of best practice for earthquake emergencies. Among them, Immersive Virtual Reality Serious Games (IVR SGs) have been suggested as an effective tool for emergency training. The notion of IVR SGs is consistent with the concept of problem-based gaming (PBG), where trainees interact with games in a loop of forming a playing strategy, applying the strategy, observing consequences, and making reflection. PBG triggers reflection-on-action, enabling trainees to reform perceptions and establish knowledge after making a response to a scenario. However, in the literature of PBG, little effort has been made for trainees to reflect while they are making a response (i.e., reflection-in-action) in a scenario. In addition, trainees do not have the possibility to adjust their responses and reshape their behaviors according to their reflection- in-action. In order to overcome these limitations, this study proposes a game mechanism, which integrates spiral narratives with immediate feedback, to underpin reflection-in-action and reflective redo in PBG. An IVR SG training system suited to earthquake emergency training was developed, incorporating the proposed game mechanism. A controlled experiment with 99 university students and staff was conducted. Participants were divided into three groups,with three interventions tested: a spiral narrated IVR SG, a linear narrated IVR SG, and a leaflet. Both narrated IVR SGs were effective in terms of immediate knowledge gain and self-efficacy improvement. However, challenges and opportunities for future research have been suggested.
|
Bellamy, L.A., Henning, T.F.P., Amor, R., Jones, D., Pancholy, P., Preston, G., van Zyl, J.E. (2022) Data Strategies for Improving Infrastructure Value and Performance in New Zealand, Smart Infrastructure and Construction (Proceedings of the ICE), https://doi.org/10.1680/jsmic.22.00008. |
Abstract: This paper reviews three digitalisation projects in the Building Innovation Partnership, an industry-led research and innovation initiative based at the University of Canterbury, New Zealand, focused on improving the transfer of infrastructure information within and between organisations. The building information modelling (BIM) to building asset management project focuses on the use of client-facing information managers and asset information specification tools to improve information transfer between building designers, contractors and facility/asset managers. The Infrastructure Asset Data project focuses on the use of metadata standards to federate and analyse data on three waters pipe networks managed by different councils. The BIM-Based Building Consenting project focuses on using BIM data to automate the assessment of building designs against the requirements of digitised building regulations and standards. Case studies associated with the three projects show that the capability of industry to transfer digital information needs to be lifted, to exploit the benefits of digital information, processes and technologies in the construction and infrastructure sectors.
|
Lim, D.M., Chau, A., Lottridge, D., Amor, R. (2022) Gesture Set Development for Augmented Reality Building Site Consenting, Proceedings of ConVR, Seoul, South Korea, 16-19 November, ISBN 978-0-9927161-4-1, pp. 336-347. |
Abstract: This research aimed to develop an intuitive gesture set for consenting officers to inspect a building on-site using an augmented reality (AR) system. When constructing a building, consenting officers are required to check a building site to ensure that what is constructed aligns with the requirements of the building. AR is the superimposition of virtual information on the user's view of the physical world. AR can enhance the building consent process by superimposing the building's information on the building site itself or facilitating the consenting officer's procedures. This requires the development of a suitable gesture set to allow users to control the system. To create the gesture set, a set of functions for the AR system was generated, then the gestures for each function were elicited from users. The gesture set was evaluated through user testing. An AR application that could recognise the gestures and trigger the functions was implemented to facilitate the user tests, and the gesture set was evaluated based on intuitiveness, learnability, memorability, and ease of performance. The results indicate that the gesture set is indeed intuitive, but further research is needed for a conclusive evaluation of the learnability, memorability, and ease of performance.
|
Fuchs, S., Dimyadi, J., Witbrock, M., Amor, R. (2022) Training on Digitised Building Regulations for Automated Rule Extraction, Proceedings of EC-PPM, Trondheim, Norway, 14-16 September, pp. 428-435. |
Abstract: We investigate automated rule extraction from building regulations using a neural semantic parser. This task is regarded as a main requirement to enable automated compliance checking in the built environment. The performance of deep learning models is strongly dependent on the quantity and quality of the training data and the task complexity, which is particularly relevant for domain-specific tasks with limited data and domain-specific terminology. The neural semantic parser is trained using a corpus of LegalRuleML rules, which were manually encoded in previous research. We identify the primary error sources for automated parsing, investigate the importance of data quality and consistency and rework the entire corpus accordingly. Extensive experiments indicate the impact of different inconsistencies. Value conditioning was evaluated to limit the effect of varying granularity, complex expressions, and tacit knowledge. Finally, we draw conclusions about the encoding guidelines and processes from a natural language processing perspective.
|
Shastri, Y., Hoda, R. and Amor, R. (2022) Spearheading agile: the role of the scrum master in agile projects, Proceedings of XP 2022, Copenhagen, Denmark, 13-17 June. |
Abstract: Scrum innovated the role of the scrum master in software engineering. The scrum master is envisioned in agile literature as the "servant leader" who serves the team in a multitude of different ways, which include promoting scrum, facilitating the team's functioning, and removing obstacles. However, empirical studies focusing on the role of the scrum master in practice are scarce. To address this gap, a Grounded Theory study with a mixed methods approach was carried out which included semi-structured interviews with 39 software practitioners and a questionnaire with 47 respondents. In this study, we present and describe the scrum master's role in agile projects in terms of (a) the grounded theory of the role of the scrum master which involves everyday activities of facilitating, mentoring, negotiating, process adapting, coordinating, and protecting; (b) the varying involvement of the scrum master in selected agile practices carried out by the team; and (c) a positive association between the presence of the scrum master and the frequency with which agile practices are carried out by the team. This study presents for the first time a multifaceted study of the multiple dimensions of the scrum master role and will enable practitioners to better manage expectations of this role in practice.
|
Wang, Q., Zhao, K., Amor, R., Liu, B., Wang, R. (2022) D2GCLF: Document-to-Graph Classifier for Legal Document Classification, Proceedings of NAACL 2022, Dublin, Ireland, 26 May. |
Abstract: Legal document classification is an essential task in law intelligence to automate the labor-intensive law case filing process. Unlike traditional document classification problems, legal documents should be classified by reasons and facts instead of topics. We propose a Document-to-Graph Classifier (D2GCLF), which extracts facts as relations between key participants in the law case and represents a legal document with four relation graphs. Each graph is responsible for capturing different relations between the litigation participants. We further develop a graph attention network on top of the four relation graphs to classify the legal documents. Experiments on a real-world legal document dataset show that D2GCLF outperforms the state-ofthe- art methods in terms of accuracy.
|
Fuchs, S., Witbrock, M., Dimyadi, J., Amor, R. (2022) Neural Semantic Parsing of Building Regulations for Compliance Checking, Proceedings of CIB W78, Melbourne, Australia, 27-29 June, IOP Conf. Ser.: Earth Environ. Sci. 1101 092022, https://doi.org/10.1088/1755-1315/1101/9/092022. |
Abstract: Computerising building regulations to allow reasoning is one of the main challenges in automated compliance checking (ACC) in the built environment. While there has been a long history of translating regulations manually, in recent years, natural language processing (NLP) has been used to support or automate this task. While rule- and ontology-based information extraction and transformation approaches have achieved accurate translations for narrow domains and specific regulation types, machine learning (ML) promises increased scalability and adaptability to new regulation styles. Since ML usually requires a large number of annotated examples as training data, we take advantage of the long history of building code computerisation and use a corpus of manually translated regulations to train a transformer-based encoder-decoder model. Given a relatively small corpus, the model learns to predict the logical structure and extracts entities and relations reasonably well. While the translation quality is not adequate to fully automate the process, the model shows the potential to serve as an autocompletion system and to identify manually translated regulations that need to be reviewed.
|
Dimyadi, J., Amor, R. (2022) Computable Planning Rules for Automating Urban Land Development, Proceedings of NZBERS, Auckland, New Zealand, 17-18 February, pp. 247-255. |
Abstract: Land redevelopment is not a trivial process and involves manually and painstakingly working with planning rules to develop various compliant options. The process is costly, inefficient, and lends itself to automation as long as we have the right computable data. Existing tools have either hard-coded rules that are not responsive to changes in the operative version of planning rules, or depend heavily on user interpretation and inputs for most of the parameters needed for the calculations. The current research combines officially published digital land information, a computable representation of planning rules, and an industry-standard financial model to provide viable compliant subdivision scenarios. Rules from the Auckland Unitary Plan and Waikato District Plans have been translated into the open legal knowledge interchange standard known as LegalRuleML, using a semi-automated method that also incorporates a robust quality assurance process. The findings suggest that the outcome of the automated process provides a close match with the result of the manual calculations. More importantly, however, the calculations can be performed very efficiently and accurately. Computable planning rules can support an automated land development evaluation process and enable a cost effective feasibility assessment for various compliant subdivision options. To test the viability of the approach, a financial model for land subdivision calculations constrained by planning rules has been incorporated into the ACALand Solution of the ACABIM commercial software platform. Several sites from various parts of Auckland and Waikato regions in New Zealand have been selected to validate the calculations. A comparison between the output of the automated calculations and those produced manually by a land development professional for a real project has been undertaken as part of the validation.
|
Guo, B.H.W., Zhang, Z., Amor, R. (2022) Applications of Building Information Modelling to Construction Health and Safety: A Systematic Review, BIMSafe Report, University of Canterbury, 68pp. |
Abstract: Background: This systematic literature review was conducted for the BIMSafe NZ project, which is a three-year, $1.7 million partnership between the construction industry and the government that intends to lower New Zealand's accident and injury rates by improving risk understanding, communication, and mitigation. Objectives: The literature review aims to understand the current status of BIM applications for construction health and safety and identify BIM-based health and safety best practices in the project lifecycle (i.e., design, procurement, and construction). It covers academic publications, related BIM standards, industry best practices, and case studies. In specific, this review study focuses on the following main aspects: ? BIM and BIM-related technologies for safety in design (SiD) and site construction health and safety management, ? Integrating BIM and health and safety into procurement processes, ? BIM information flow and management for whole lifecycle health and safety management, ? Drivers and barriers to the implementation of BIM for construction health and safety management, ? Existing standards, guidelines, and best practices of BIM-based health and safety. Methods: A systematic review method was adopted to identify and analyse relevant academic articles, BIM for safety standards, guidelines, best practices, and industry case studies. The method consists of five main steps: literature search, selection, coding, data analysis, and discussion. BIM-based safety in design: Previous research efforts were focused on developing BIM-based rule checking and risk assessment systems. Several knowledge bases have been developed to support rule-based reasoning and risk assessment processes. In these research efforts, BIM is mainly used as a database, and visualization platform from which objects and their attributes were extracted as inputs for reasoning and assessment, and outputs (e.g., hazards and control measures) are displayed to facilitate communication. Despite these advancements, there is a strong need to develop a comprehensive knowledge base that enables identifying most of, if not all, hazards, especially those caused by the spatial and temporal relationships among building elements and construction activities. Second, how these techniques work through different design phases (i.e., concept, developed, and detailed design) needs to be investigated. BIM-based health and safety in procurement: There is minimal research investigating BIM-based health and safety during the procurement stage. Nevertheless, BIM4H&S Working Group developed a Project Information Requirements (PIR) template to help the client specify health and safety information requirements in a BIM project. These requirements were set up based on the PAS 1192-6 and classified into three types: functional, role and process, and information cycle requirements. BIM-based site health and safety management: Significant research achievements have been made in several aspects of BIM-based site health and safety management, including hazard management, risk assessment, site safety inspection and monitoring, temporary structure, heavy equipment, safety training and education, and work space planning. Future research directions: This report recommends six specific future research directions: ? Develop safety object libraries ? Develop a comprehensive computable safety knowledge base ? Standardize health and safety information structure ? Evaluate the performance of BIM-based health and safety applications by accident reduction ? Investigate BIM-based multi-stakeholder collaboration on health and safety ? Integrate health and safety into the NZ BIM Handbook