Most decision making in construction projects is accomplished by human beings based on manually collected information. This process is labor-intensive, error-prone, and heavily reliant on the knowledge and experience of the decision makers as well as the quality and reliability of the available information. Cyber-physical systems (CPS) offer the potential to transform traditional data collection approach and strengthen the coordination between cyber models and physical assets in a construction project. This paper describes a knowledge-based CPS architecture that aims at augmenting human decision making in construction. Domain knowledge, such as site policies, safety rules, and workflow logic, is elicited and compiled in an ontology-based knowledge hub. With the support of bi-directional cyber-physical communication, the knowledge hub inquiries relevant data from cyber models and physical assets, integrates real-time project data, and synthesizes context-specific information to facilitate human decision-making. This paper outlines the key challenges and technical requirements for creating the knowledge hub and its linkages with traditional CPS systems. It also presents two deployment scenarios in highway and building construction projects to demonstrate the realization a knowledge-based cyber-physical system and its potential in facilitating coordination and improving project performances in safety and efficiency.