Deep learning aided fingerprint based beam alignment

1. A method implemented by a network node, the method comprising:

obtaining input data comprising a wireless transmit/receive unit (WTRU) location, a number of WTRU, and a desired received signal strength;

processing the input data with a trained neural network to generate a set of one or more beam-pairs, wherein each beam-pair is associated with the WTRU location, the number of WTRUs, and the desired received signal strength;

transmitting, to a WTRU, information indicating one more candidate beam-pairs, wherein the one of more candidate beam-pairs are based on the generated set of beam-pairs;

transmitting, for at least one candidate beam-pair of the one or more candidate beam-pairs, a signal using a it beam of the at least one candidate beam-pair, and

receiving, from the WTRU, information indicating at least one suitable candidate beam-pair from the one or more candidate beam-pairs, wherein the at least one suitable candidate beam-pair provides the desired received signal strength.

2. The method of claim 1 , wherein the WTRU location is received from the WTRU based on an initial network access process.

3. The method of claim 1 , wherein the number of WTRUS is known to the network node according to a vehicular density.

4. The method of claim 1 , wherein the generated set of one or more beam-pairs is generated using a softmax algorithm.

5. The method of claim 4 , wherein the softmax algorithm generates probabilities associated with sets of beam-pairs.

6. The method of claim 1 , wherein the generated set of one or more beam-pairs is stored in a database during a training phase.

7. The method of claim 6 , wherein the trained neural network uses weights determined from the training phase, and wherein the training phase comprises:

obtaining training samples for each training location;

initializing weight vectors as random values; and

iteratively performing the following steps until a convergence metric threshold is reached:

computing neuron outputs for each layer using the respective weight vector;

applying a softmax function to obtain class probabilities;

calculating weight matrices and bias vectors; and

performing error backpropagation.

8. The method of claim 1 , wherein processing the input data with the trained neural network comprises:

obtaining a plurality of sets of one of more beam-pairs for different vehicular densities at the WTRU location; and

using the trained neural network coupled with a softmax classifier to generate the set of one of more beam-pairs from the plurality of sets of one of more beam-pairs based on the vehicular densities at the WTRU location, wherein using the trained neural network uses the weights determined from a training phase.

9. The method of claim 8 , wherein the vehicular densities comprise the number of WTRUs at the WTRU location.

10. The method of claim 8 , wherein the trained neural network is a deep learning feedforward neural network.

11. The method of claim 1 , wherein the trained neural network uses weights determined from a training phase, the method further comprising:

obtaining training samples for at least one training location;

initializing weight vectors as random values; and

iteratively performing the following steps until a convergence metric threshold is reached:

computing neuron outputs for at least one layer of the trained neural network using the respective weight vector;

applying a softmax function to an output layer of the trained neural network to obtain class probabilities;

updating weight matrices and bias vectors; and

performing error backpropagation.

12. The method of claim 11 , further comprising:

determining a loss function between a predicted class probability and a true class probability, wherein the convergence metric threshold is reached if the loss function is less than the convergence metric threshold.

13. The method of claim 1 , further comprising:

selecting the at least one suitable candidate beam-pair to transmit data to a receiver.

14. The method of claim 13 , wherein selecting the at least one suitable candidate beam-pair comprises using a multi-functional beam transmission scheme.

15. The method of claim 13 , wherein selecting the at least one suitable candidate beam-pair is repeated periodically.

16. The method of claim 13 , wherein selecting the at least one suitable candidate beam-pair is performed upon detection of a triggering event, wherein the triggering event is a change in a parameter of the WTRU.

17. The method of claim 1 , wherein the network node is a base station.

18. A network node comprising:

a processor; and

a non-transitory computer-readable medium storing instructions operative, when executed by the processor, to:

obtain input data comprising a wireless transmit/receive unit (WTRU) location, a number of WTRUs, and a desired received signal strength;

process the input data with a trained neural network to generate a set of one or more beam-pairs, wherein each beam-pair is associated with the WTRU location, the number of WTRUs, and the desired received signal strength;

transmitting, to a WTRU, information indicating one or more candidate beam-pairs, wherein the one or more candidate beam-pairs are based on the generated set of beam-pairs;

transmit, for at least one candidate beam-pair of the one or more candidate beam-pairs, a signal using a transmit beam of the at least one candidate beam-pair; and

receive, from the WTRU, information indicating at least one suitable candidate beam-pair from the one or more candidate beam-pairs, wherein the at least one suitable candidate beam-pair provides the desired received signal strength.