Review of key technologies of MAC layer in radio network

Cognitive radio (CR) technology has been paid more and more attention by the way of "two utilization" of the authorized spectrum, which effectively relieves the contradiction between the lack of spectrum resources and the increasing demand for wireless access.

In order to realize CR users ' use of spectrum holes while avoiding harmful interference to authorized users, the media access control (MAC) layer of CR network not only provides traditional services, such as media access control and robust data transmission, but also supports a new set of functions, That is, the effective realization of the opportunity spectrum utilization without interfering with the authorized user. These new functions are embodied in the spectrum detection management, access control, dynamic spectrum allocation, security mechanism and cross layer design of MAC layer, and the above technologies are discussed separately.

1 MAC Layer Spectrum detection management

The MAC layer spectrum detection management is mainly used to control the implementation of the physical layer spectrum detection algorithm, such as deciding which channels to detect and when to detect. At present, the research of the MAC layer Spectrum detection management mainly focuses on the selection and optimization of detection strategy and detection parameters, including the selection of detection mode, the setting of detection period and length, the selection of detection channel and the setting of detecting the silent period.

According to the different timing of CR user detection, the detection mode can be divided into cycle detection and on-demand detection. Cycle detection means that CR users detect the channel according to a certain period of time, not only when the data is sent. In this way, the state information of the channel can be collected periodically, which is beneficial to estimating the channel state and locating the spectrum hole quickly. On-demand detection is when CR users have data to send to detect the channel. Compared with the cycle detection, on-demand detection reduces the unnecessary detection overhead, but detects the frequency of the spectrum hole for a long time. According to the principle of "energy efficiency", the adaptive selection of Detection mode is realized by the tradeoff between the energy consumed by the detection and the delay of finding the idle channel.

In the cycle detection, the selection of the appropriate detection cycle is critical, if the detection cycle is too large, will be unable to detect some of the free spectrum, and loss of some access opportunities, but also due to the failure to detect the presence of authorized users to produce harmful interference, if the detection cycle is too small, it will lead to too frequent detection, consumption of unnecessary energy. Cycle detection usually includes two mechanisms of synchronous cycle detection and asynchronous periodic detection. Synchronization cycle detection mechanism for all channels to set the same detection start time and the same detection cycle, the implementation of simple, lack of flexibility, corresponding to the asynchronous cycle detection mechanism flexibility has received more attention.

In the paper [1], an adaptive asynchronous detection period optimization algorithm is proposed with minimizing loss access opportunity as the optimal target. The algorithm is adaptive to each channel to set its detection cycle, to reduce the length of the location of idle spectrum, maximize the use of access opportunities played a role, but for each channel, the detection cycle is still a fixed interval, that is, the optimal detection cycle once selected will no longer change, is still essentially a detection mechanism based on the fixed detection cycle.

As the extension of the fixed-period detection mechanism, this paper presents a detection mechanism based on variable detection period (FSP), introduces the "Detection interval control factor", and adjusts the factor to reduce the detection period in the region where the channel State may change, improve the detection efficiency and reflect the flexibility of the cycle change. In order to further extend the variable detection cycle of FSP mechanism to the stochastic detection cycle, a random detection mechanism (RAPSS) is introduced, and a more general detection period Optimization Model (MRM-SPO) is proposed. At the same time, the existence of detection error caused by the limitation of Physical layer detection algorithm in practical application is considered, and the influence of the CR user Latency channel mechanism, which is introduced by the spectrum collision of authorized users, to avoid the detection cycle optimization is discussed.

As another main parameter of the cycle detection, the detection time length is the embodiment of the test quality and the measurement speed compromise in nature. Shortening the length of the test will result in a decrease in the quality of the test, while increasing the length of the test can improve the quality of the test, but also reduce the utilization of available free spectrum. The length of detection is directly related to the underlying hardware and the physical layer detection algorithm, and can be selected and optimized according to the utilization rate of the free spectrum, the compromise of detection speed and the performance of the CR users.

In order to find the spectrum opportunity quickly, besides optimizing the length of detection, it involves the selection of Detection channel. The existing research on the selection of detection channels mainly includes the following aspects: Selecting the most likely idle channel for detection, optimizing the sequence of channel detection.

In addition, the detection of the silent period is also an important research content of MAC layer spectrum detection management. From the point of view of CR network, when CR users detect a spectrum within the CR network, all other users in the system who work on the spectrum should remain silent for a period of time to ensure that the CR user's own communication does not interfere with the detection of the authorized user's signal, which is known as the silent period.

The silent period can be divided into two kinds: synchronous silent period and asynchronous silent period according to different ways of realization. In the synchronous silent period, at the same time CR users stop transmitting signals on all available channels of the system, the setting is simpler, each CR user can detect all channels; asynchronous silent period is one or more CR users stop transmitting signals on a specific channel occupied, each channel can have different quiet periods, Cr users typically only detect the channel they occupy. Because the realization of synchronous quiet period needs the support of dynamic wideband filter, the current research pays much attention to the asynchronous silent period, which includes the asynchronous silent period with the protection interval and the asynchronous silent period without overlapping time.