Corresponding Author

Wen-Hsin Wang

Document Type



AIoT solution based on the AI (Artificial Intelligent) and IoT (Internet of Things) is considered state-of-the-art technology and has emerged in various business environments. To enhance intelligent traffic quality, maximize energy saving and reduce carbon emission, this study applied an AIoT technology based on traffic counting modules and people behavior modules as traffic inference systems. Applications of the IoT technology based on WiFi, 3G/4G and NB-IoT (Narrowband IoT) was conducted gradually in key demonstration roads and cities worldwide, and the development and evaluation results were aligned to an action research framework. The five phases in the action research included designing, collecting data, analyzing data, communicating outcome, and acting phases. During the first two phases, problems of functional operations in traffic were verified and designed for network services by ICT (Information and Communication Technology) and IoT technologies to collection traffic big data. In the third phase, stakeholders may use basic statistic or further deep learning methods to solve traffic scheduling, order and road safety issues. During the fourth and fifth phases, the roles and benefits of stakeholders participating in the service models were evaluated, and issues and knowledge of the whole application process were respectively derived and summarized from technological, economic, social and legal perspectives. From an action research approach, AIoT-based intelligent traffic solutions were developed and verified and it enables MOTC (Ministry of Transportation and Communications) and stakeholders to acquire traffic big data for optimizing traffic condition in technology enforcement. With its implementation, it will ultimately be able to go one step closer to smart city vision. The derived service models could provide stakeholders, drivers and citizens more enhanced traffic services and improve policies’ work more efficiency and effectiveness.