How to build a TDNN with two data sequence inputs

Hello,

I am configuration a TDNN for the communication data. Architecture as below:

There are two data sequences, I and Q, to be fed into the network. I tried timedelaynet. The network, however, seems to accept only one data sequence. Anyone can tell me how to input two parellel data sequence into the timedelaynet? Thanks.

example as below;

net = timedelaynet(1:10,50);

[Xs,Xi,Ai,Ts] = preparets(net,X,T);

net = train(net,Xs,Ts,Xi,Ai);

Answer 1

[Aditya]

Hello Yi,

To input two parallel data sequences into a ‘timedelaynet’, you can concatenate the two sequences and treat them as a single input sequence. Here's an example of how you can modify your code to achieve this:

X = [I; Q]; 

This modified input sequence can be used with the ‘preparets’ function to prepare the data for training the ‘timedelaynet’.

For more detailed information on the ‘timedelaynet’ function, you can refer to the documentation provided by MathWorks: https://in.mathworks.com/help/deeplearning/ref/timedelaynet.html

Hope this helps!