Radius selection for lattice sphere decoder-based block data transmission systems

Abstract

Lattice sphere decoder (LSD) searches lattice points in space within a certain radius (d), where the closest point obtained is considered the solution. It is well known in LSD, when the initial radius (d) increases, the complexity increase. Therefore, this paper aims to obtain an initial radius (d) exact expression to reduce the system complexity with reasonable performance. The derived expression shows that initial radius (d) depends on lattice dimension n, signal-to-noise ratio (c), and noise variance r 2 . Hence, this paper proposed a new LSD for BDTS based on this initial radius technique. The proposed LSD achieves a good balance between complexity and performance. Other analytical expressions for complexity and performance in relation with (d) are also derived. It has been observed that the convergence between the analytical system performance and complexity with their respective simulation results.

Massive MIMO Detection Techniques: A Survey

“…The main idea behind the sphere decoder (SD) algorithm is to search only through the constellation points that are confined within a sphere with a predetermined radius "d" [224] [225]. Figure 9 shows the geometrical representation of the SD algorithm where the small blue nodes represent all possible transmitted symbols.…”

Nested Variational Chain and Its Application in Massive MIMO Detection for High-Order Constellations

“…Many research studies have been carried out for signal detection in massive MIMO systems [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. It is well known that the maximum likelihood detection presents the best detection performance with the cost of exponentially growing computational burden [3].…”

“…The sub-space searching-based category originates from the idea of reducing the searching space of all possible lattice points with unacceptable complexity. Sphere decoding tries to replicate the maximum likelihood performance by diminishing the searching space, the dimension of which grows up with the number of antennas as well as the modulation order, making it prohibitive for the large-scale or high-order MIMO systems [6,7]. Another two local searching-based approaches were proposed by the name of likelihood ascending search and reactive tabu search [8][9][10], and the basic idea behind them is to search through a proximity sub-space around a given initial solution.…”

Sphere decoding algorithm for multiuser detection in a distributed antenna system

“…It is shown that SD receivers give a performance close to ML receivers for different users' positions within a cell. The search complexity of the SD can be reduced with an appropriate selection of the initial search radius, as shown in [17]. However, a fixed value of the initial radius will increase the search complexity.…”