| 摘要: | 研究期間：10205~10304;Recently, due to the advances in wireless communication, MEMS, VLSI technology, small equipment with capabilities of communication, sensing, computation, and storage gradually becomes feasible. In addition, the powerful cloud computing enables various amazing applications through internet. Hence, based on these factors, the concept of “Internet-of-Things” (IoT) is developed. In the scope of “IoT”, besides the conventional communications between humans, more links are introduced, namely, human to machine (H2M), machine to human (M2H) and machine to machine (M2M). On the other hand, the data is not simply delivered between machine nodes. Because the computation capability can be upgraded in each machine node, simple and preliminary decision and analysis are performed to bring the feature of Intelligence in IoT. This project develops the communication techniques of spectrum cognition and compressed sensing for intelligent sensor networks. To solve the communication problems from limited spectrum resources and to enhance the spectrum utilization, we aim to use the technique of spectrum cognition that utilizes the spectrum as a secondary user without interfering the primary licensed users. Furthermore, it also can reduce the interference problem for numerous nodes in M2M networks. Secondly, the major power consumption in the sensor network comes from the communications. Hence, we plan to design a compressed sensing engine that can reduce the samples from various BIO or MEMS sensors for transmission. Due to the characteristics of ad hoc network, we tend to rely on distributed algorithms. Next, the relationship between spectrum efficiency and energy efficiency is going to be surveyed so that the power consumption for signal transmission can be suppressed aggressively by a suitable modulation scheme. The dynamic voltage and frequency scaling (DVFS) is going to be adopted to achieve the goal of a low power design. To convey the compressed image data, the transmission rate is targeted at 2 Mbps, a little bit higher than conventional wireless sensor networks. And the operation modes can be classified as sleep mode, low duty mode and standard mode. Finally, this communication digital signal processor together with the smart video processor, smart sensor interface, power and clock management, wireless powering and various sensors including BIO sensors, MEMS sensors and image sensors are integrated with the help of design automation tools. We aim to accomplish the design of a sensing system in this intelligent sensor network for life and care. |
