Framework Optimization And Power Control Research In Cognitive Radio

The contradiction between high scarcity of unlicensed spectrum resources and inefficiency usage of licensed spectrum resources needs to be resolved radically, which demands emerging a new technique imminently. On the other hand, in cognitive radio systems, a cognitive radio user (cognitive user, CU) can access licensed spectrum holes opportunistically in dynamic spectrum access manner, meanwhile which does not affect the normal communication of a licensed user (primary user, PU) can improve the utilization efficiency of the licensed spectrum effectively and resolve the problem of the scarcity of spectrum resources. Detecting spectrum holes is the foremost task of the cognitive user. Furthermore, in the process of using the spectrum holes the cognitive user is not allowed to interfere with the primary user. Thus the cognitive user must carry spectrum sensing rapidly and accurately.At present, there are several common spectrum detection algorithms. Their principles, advantages and disadvantages are analyzed in this paper. The principle of energy detection is that the cognitive user accumulates energy on interesting spectrum and observational time. When the accumulating energy value (detected energy value) is larger than the preset threshold, the cognitive user decides that the primary user is present; otherwise, the cognitive user decides that the primary user is absent. Since energy detection has the advantages in simplexes and easy operation, it has the most practical use and become the most common detection algorithm.In most of current studies, the cognitive user makes a simple binary decision whether the primary user is present based on the detected energy value and the decision threshold. Afterwards, the cognitive user adopts fixed transmitting duration or waiting duration usually. Be similar to the soft decision in communication systems, we can foresee that the simple binary decision will cause some usable information to be lost since the detected energy values cover the usable information whether the primary user is present. Based on above-mentioned consideration, this paper proposes a new framework model, i.e., utilizing the detected energy information adjusts the transmitting duration and waiting duration adaptively, more to the point, taking normalized channel efficiency and lost performance to be the optimization indexes we give the optimal transmitting duration and waiting duration respectively. The concrete optimization course is summarized as follows. For a detected energy value which locates a energy interval, assumed that the energy interval corresponding optimal transmitting duration or waiting duration is an unknown quantity, we compute system performance (normalized channel efficiency or lost performance) brought by the unknown quantity firstly. Then, on condition that normalized interference duration is limited, in order to maximize the normalized channel efficiency, we compute the optimal transmitting duration; in order to minimize the lost performance, we compute the optimal waiting duration so as to achieve the tradeoff between sensing overhead and lost idle spectrum duration. Finally, making the width of each energy interval approach to zero, the optimal transmitting duration and waiting duration derived from the front step can get the limit value.The magnitude of the cognitive user's transmitting power affects the cognitive radio system's performance (average throughput), and also exerts a direct influence to the primary user's interference level (peak interference power). Thus, controlling transmitting power to achieve a tradeoff between the average throughput and the peak interference power is necessary. In AWGN channel, this paper lets the detected energy value as a variable and the tradeoff between the average throughput and the peak interference power as an object, the adaptive optimal transmitting power value of the cognitive user is given.Finally, simulation experiments by MATLAB validate the scheme proposed by this paper. The numerical results show that the proposed scheme has better performances compared with the conventional scheme. With the accuracy of dividing the detected energy value (i.e., the width of energy intervals is closer to zero) improving, the adaptive model proposed by this paper can bring more performance improvement.