低工作週期機器對機器網路中時間嚴格之信息傳輸

Abstract

M2M communication is an enabling solution for many applications, including industrial control and smart buildings. Take the temperature monitoring system for machinery in a factory as an example. The temperature sensors send back the readings to a central controller. The controller will shut down the system if the readings are over a predefined threshold. In smart buildings, lighting control systems consist of different sensors such as light sensors and motion sensors. These sensors detect the behaviors of human beings and turn on/off the lights accordingly. A common problem among the above applications is that many machines, especially the sensors, are usually battery-powered. Therefore, they have extremely limited energy budget. Given that it is also very inconvenient to replace or recharge the battery, how to maintain an M2M network in an energy-saving manner becomes a critical issue in the design of M2M networks. Turning the radio off whenever possible (i.e., duty cycling the machines) is one of the most common methods for energy management. Duty-cycle control, however, results in a longer latency of message dissemination. Although longer latency is not a serious problem for non-time critical messages such as regular temperature readings, it will lead to devastating results for time-critical messages such as fire alarms. Most of the existing researches focused on minimizing the latency of non-time critical messages or time-critical messages seperatedly. A joint design to make tradeoff between the latency of these two types of messages is still missing. In this thesis, a set of specially designed sequences that control the duty cycle of machines is developed to make a better tradeoff. Our simulation results show that the end-to-end latency of time-critical messages is shortened by 13% to 42% depending on the duty cycle, while that of non-time critical message is only increased by less than 7%. The proposed solution is also implemented in an IEEE 802.15.4 network. The result shows that our sequences can prolong the lifetime of machines by 22% to 333% and reduce the end-to-end delay by 26%.