| Summary: | 碩士 === 國立臺灣大學 === 電信工程學研究所 === 106 === With the rise of Industry 4.0, various kinds of sensors, data collection devices, and data analysis devices in the factory environment are now quickly developed toward high degree factory automation. In such factory environment, there must be many devices that make use of wireless transmission technology for their data transmissions. However, due to limited spectrum resources, there can be serious channel contention and interference behavior among these wireless devices. How to control and manage these wireless devices in the factory is the key part of this research. In this thesis, the types of devices are assumed to fall into two main categories. The first one is the high throughput device and data collection device with requirement of high throughput. The second is the delay sensitive device with strict requirement on packet delay time. The algorithm designed through this research will attempt to allow all devices to meet their QoS requirements as much as possible among the aforementioned constraints.
In this thesis, a novel approach called the Proactive QoS-aware Device Access Control Method(PQDA) is proposed. The PQDA method includes three major phases. i.) The AP frequency selection phase, ii.) device transmit target selection phase, and iii.) The device control phase. In the first phase, by modifying the Welsh-Powell Algorithm, the APs in the factory can avoid co-channel or near-band interferences, in order to maximize the overall wireless network capacity. In the second phase, by estimating the throughput of devices, and design rules derived from the simulation, we develop an approach to maximize the number of devices meeting their QoS requirements. In the last phase, while the factory is actually working, the factory can respond in advance, one step earlier to reduce the occurrence of disturbance by estimating the impact of devices before joining the wireless network.
In the final part of this study, the feasibility of PQDA and its advantages are verified by simulations, and the results are compared with the performance obtained via common methods in traditional WiFi network management and planning approaches, proving that the PQDA method can be adapted in various kinds of factory environments, and enable most devices to meet their QoS requirements.
|
|---|
