Cognitive-radio-networks with multiple input multiple output is an effective method to improve spectrum efficiency. Linear pre-coders is a crucial technology to realize this goal. However, most of existing algorithm developed based on the Gaussian input assumption. This assumption limited the application of linear precoding and caused great performance loss. In this paper, for such a problem, a linear pre-coders algorithm for maximizing the secondary users spectrum efficiency is designed. Different from previous spectrum efficient algorithm which is developed under the Gaussian input assumption, the proposed algorithm can work under any input assumption by inheriting the advantage of mercury water filling algorithm (MWF) and singular value decomposition (SVD). Numerical results show that the proposed algorithm exhibits a large gain over existing algorithms when we take a practical modulation type.

In this study, the authors mainly study the degrees of freedom (DoFs) of M × N user single input multiple output (SIMO) (or multiple input single output (MISO)) X channels, in which the channel coefficients are drawn from a continuous distribution and data‐sharing among users are unavailable. First, the authors establish an outer bound of DoFs for multiple‐in multiple‐out (MIMO) X channel, which is tighter than those in previous work. For SIMO X channel with R antennas at each user, the authors prove that the DoFs is [ MNR /( M + NR − R )], which is achieved by symbol extension and interference alignment. As the transmitters and receivers are interchangeable in the network considered, the authors obtain corresponding results for MISO X channel. Finally, the authors generalise our result to MIMO X channel with R antennas at each user, and show that its DoFs is [ MNR /( M + N − 1)].

In this paper, the problem of joint power allocation (PA) for the downlink of a multi-cell distributed antenna system (DAS) is addressed. Motivated by practical applications, we focus on a reasonable scenario that only the large-scale channel state information at the transmitter (CSIT) is available. Based on the observation that the capacity-achieving input covariance for each cell is diagonal and that PA optimization is enough to achieve the maximum ergodic sum capacity, we formulate a joint PA optimization problem to maximize the ergodic sum capacity of the system with a total transmit power constraint for each cell. A rather precise closed-form approximation of the ergodic sum capacity is then introduced and taken as the objective function instead so that the original joint PA optimization problem can be simplified. Finally, we propose an iterative PA scheme based on the simplified joint PA optimization problem, in which Signomial Programming (SP) is used. Monte Carlo simulations show that the proposed scheme converges quickly and can offer nearly optimal system ergodic sum capacity. Thus, we refer the proposed PA scheme as a suboptimal one. Moreover, from the simulations we can see that a significant performance gain can be achieved by multi-cell joint PA in DAS with only large-scale CSIT.

In this paper, we investigate the capacity of continuously time-varying multiple-input multiple-output (MIMO) systems in frequency-flat Rayleigh fading environment with perfect interleaving. By introducing the Gauss-Markov model to describe the channel variation and employing minimum mean square error (MMSE) channel estimation based on the pilots, we derive very tight lower and upper bounds of the ergodic capacity in closed-form. We also derive the optimal power allocation between the pilot and data vectors, which maximize the lower bound of the ergodic capacity. Interestingly, the optimal allocation is independent of the channel variation parameter and can be easily computed (no feedback is required). The optimal training interval can be obtained via numerical optimization. Finally, two different transmit schemes are compared via simulation. It is shown that, in fast fading or high SNR environments, the scheme with optimal power allocation almost has the same performance as the scheme with equal power allocation. However, in slowly fading or low SNR environments, the former has much better performance than the latter.

This paper presents frequency-domain equalization (FDE) using iterative channel estimation for high-speed transmission with cyclic-prefix multicarrier direct sequence spread spectrum CDMA (MC-DS-CDMA). Cyclic-prefix DS-CDMA simplifies many issues arising in receiver due to the high-speed transmission, such as delay-spread of the channel. It also reduces the complexity of the equalizer and enables the equalization in frequency domain. With pilot-aided channel estimation and OFDM-style cyclic prefixes, a frequency-domain MMSE equalizer achieves the suppression of multipath interference. Yet the accuracy of channel estimation restricts the performance of the equalizer. In order to acquire more accurate channel estimation, an iterative algorithm is presented in this paper. Simulation results show, with the employment of this algorithm, the performance of the frequency-domain equalizer has a significant improvement. Compared to the original RAKE receiver with severe interference, after 5 iterations, the SNR of despread signals is enhanced by about 13 dB. This higher SNR can satisfy the requirement of demodulator and decoder, and support the full usage of Walsh codes and high order modulation as well.