Power Allocation and Beam Scheduling for Multi-User Massive MIMO Secret Key Generation
Power Allocation and Beam Scheduling for Multi-User Massive MIMO Secret Key Generation
复制标题
多用户大规模 MIMO 密钥生成的功率分配和波束调度
DOI：
10.1109/access.2020.3022052
复制
发表时间：
2020-01-01
期刊：
影响因子：
3.9
通讯作者：
Li, Guyue
中科院分区：
文献类型：
--
作者：
Sun, Chen;Li, Guyue
在多用户海量多输入多输出（MIMO）通信中，密钥生成存在计算复杂度高、导频开销大以及海量MIMO信道尺寸大造成的用户间干扰等问题。为了解决这些问题，本文提出了一种利用波束域信道进行信道特征提取的密钥生成方法。该方法通过设计良好的功率分配和波束调度算法，有效地降低了信道特征维数，减小了用户间干扰。具体而言，建立了多用户密钥生成模型，并将信道特征提取矩阵的设计问题表述为求和密钥率最大化问题。为了解决这个问题，我们导出了密钥率的上界，并证明上界最大化是一个凸优化。解决上界极大化问题后，基站得到所有用户的和波束。为了将选择的波束分配给每个用户，提出了一种高效的波束调度算法，该算法收敛于计算复杂度低的局部最优解。数值结果表明，在该方案中，合法用户的比特不一致比（BDR）达到10−2，而窃听者的比特不一致比趋于0.5，验证了该方案的可行性和安全性。
Secret key generation in multi-user massive multiple-input multiple-output (MIMO) communication suffers from high computation complexity, pilot overhead, and inter-user interference caused by the large dimension of the massive MIMO channel. To address these problems, this paper proposes a novel key generation approach that exploits the beam domain channel for channel feature extraction. The proposed approach can effectively reduce the channel feature dimension and mitigate the inter-user interference, via the well-designed power allocation and beam scheduling algorithms. Specifically, the multi-user key generation model is built and the design problem of channel feature extraction matrices is formulated as a sum key rate maximization problem. To tackle this problem, we derive an upper bound of the secret key rate and prove that the upper bound maximization is a convex optimization. After solving the upper bound maximization, the base station (BS) obtains the sum beams for all the users. To allocate the selected beams to each user, an efficient beam scheduling algorithm is proposed, which converges to a locally optimal solution with low computational complexity. Numerical results illustrate that in the proposed scheme, the bit disagreement ratio (BDR) of legitimate users achieves 10−2, while that of the eavesdropper tends to 0.5, confirming the feasibility and security of the proposed scheme.