Publication Type:

Conference Paper

Source:

8th International Conference on Computing, Communications and Networking Technologies, ICCCNT 2017, Institute of Electrical and Electronics Engineers Inc. (2017)

ISBN:

9781509030385

URL:

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85041394397&doi=10.1109%2fICCCNT.2017.8204042&partnerID=40&md5=5e29781f577c76457acbd2df8e0ca1ac

Keywords:

Assessment criteria, Classification rules, Classification technique, Concretes, Data mining, Decision trees, Identification and evaluation, Intelligent systems, Learning algorithms, Learning systems, Models, Performance parameters, Personnel training, Predictive performance models, Reinforced concrete, Reinforcement, Simulators, Support vector machines, Virtual training simulators, Vocational training

Abstract:

Construction rebars (steel concrete reinforcing bars) are used to provide structural strength and reinforcement for the concrete structure. This requires the bending and cutting of the rebar to proper size before they can be used for construction. A novel haptic based barbending simulator has been devised which enables the trainees to learn and improve the construction rebar bending skill in a safe and controlled way. With its limited assessment and reporting features, the computerised virtual training simulator proves to be effective in training. Adding the features like personalized skill tracking and predictive performance modeling holds even more potential in supporting the training program. Towards this goal, a user performance modeling needs to be done which includes an initial study on performance parameters, assessment criteria and data collection before remodeling the simulator. This paper presents a study on the performance parameters for the bar bending simulator and also evaluates its effectiveness in modeling expert and novice performances. During this study we also hypothesize the parameters that can distinguish an expert and novice performances which was validated with 2 classification techniques - Support vector machine and J48 Decision tree. While revealing the classification rules J48 algorithm provides 78.94% accuracy where as SVM provides 94.737% accuracy. The study also shows that the 2 performance parameters force applied over time and bend angle accuracy are effective to distinguish expert and novice level of expertize for the construction rebar bending skill. © 2017 IEEE.

Cite this Research Publication

B. M. Menon, Aswathi, P., Deepu, S., and Rao R. Bhavani, “Identification and evaluation of performance parameters for RE-BAR bending training simulator”, in 8th International Conference on Computing, Communications and Networking Technologies, ICCCNT 2017, 2017.