Adaptive beamforming of MIMO system using optimal steering vector with modified neural network for channel selection

Babu, P. Sekhar; Naganjaneyulu, P. V.; Prasad, Kavirayani R.

doi:10.1142/s0219691319410066

Cited by 3 publications

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The neurons in the hidden layer and the output layer are calculation nodes, and the hidden layer nodes generally use the same Radial basis function. The basis function of RBF network hidden layer unit can be Gaussian function, reflection S-line function, multiple quadratic function, etc., and the excitation function is generally Gaussian function[11].For the designed static beamforming neural network, the received signal enters the RBF network from the antenna array. The input layer to the hidden layer is a fixed connection of weight 1.…”

mentioning

confidence: 99%

A Static Beamforming Method for Neural Networks with Continuous Spatial Coverage

Yang,

Jiang,

Liu

et al. 2023

Advances in Transdisciplinary Engineering

View full text Add to dashboard Cite

The static beamforming algorithm obtains the beam response of the corresponding angle lattice within the range of scanning angles of interest by traversing different angles, resulting in the problem of discontinuous coverage in the spatial domain [1]. And the denser the airspace, the greater the beam storage resources required by the system. In response to the problem of large pointing angle estimation errors caused by discontinuous spatial coverage in traditional static digital beamforming algorithms, a beamforming algorithm based on RBF neural network is proposed. By utilizing the good generalization characteristics and nonlinear approximation ability of RBF neural network, the corresponding relationship between the array heading vector and beamforming weight is established to fit data on non lattice points, thus achieving continuous spatial coverage of static beamforming [2]-[3]. At the same time, after a large amount of precise network training, the use of neural networks can quickly obtain beamforming results in the same application scenario. Through theoretical analysis of the principle and ergodicity advantages of static beamforming algorithm based on RBF neural network, a simulation model was established, and the beamforming results of this algorithm were compared with traditional static beamforming methods. The simulation results show that the proposed static beamforming method based on RBF neural network has accurate beamforming ability and good robustness to amplitude and phase errors. The trained neural network can quickly obtain beamforming results. It has high practical application value.

show abstract

mentioning

confidence: 99%