Graphical Scripting Approach Integrated with Speech Recognition for BIM-Based Rule Checking

A building design must satisfy the requirements from diverse sources such as a building code, owner's specifications, design guidelines, and project requirements. Such requirements entail various types of rules and execution plans. In addition, the increasing complexity and requirements of a building design have required various rule-based checking platforms and capabilities. To explore a new rule-checking method, this paper proposes a node-based scripting approach with integrated speech recognition to define checking rules and evaluate the quality of BIM models for its completeness. The node-based modeling feature and its rule-based validation have a potential to define the rule components units as nodes, to offer flexibility and usability with multiple parameters, to extend rules as nodes, and to manage a large number of rules attached to a BIM design. To achieve this goal, graphical scripting, Marionette in Vectorworks, has been explored as a modular-based rule-checking approach. According to the rule types required during the early design phase, which is the work scope of this research study, the nodes were categorized by the four distinct types: object selection, rule definition, rule implementation, and validation reporting. In addition, to explore the new feature integrated with Marionette, this research study involves the investigation and development of voice recognition-based rule checking. The development and implementation of node-based rule checking features show significant potentials to provide flexible and extensible design checking rules as a form of a module.