STRLCPY/wirelesscomm

Changed m to mlx file
Sundeep Rangan committed 2 years ago

9ced758f

1 parent ad5152cc

Revision indexing in progress... (symbol navigation in revisions will be accurate after indexed)

■ ■ ■ ■ ■ ■

unit03_fading/partial/lab_chan_sounder.m

1	-	%% Lab: 5G Channel Sounding with Doppler
2	-	% Channel sounders are used to measure the channel response between a TX
3	-	% and RX. These are vital to study propagation and are also an excellent
4	-	% tool for debugging the front-end of a transceiver system. In this lab,
5	-	% we will simulate a simple channel sounder over a fading channel with
6	-	% time-variations and Doppler. The very same tools are used
7	-	% in radar.
8	-	%
9	-	% The digital communications class covered a simpler version of this lab
10	-	% with a static channel.
11	-	%
12	-	% In doing this lab, you will learn to:
13	-	%
14	-	% * Describe cluster-delay line (CDL) models
15	-	% * Get parameters for 5G CDL models using the
16	-	% <https://www.mathworks.com/products/5g.html 5G MATLAB toolbox>
17	-	% * Represent antenna orientations using global and local frames of
18	-	% reference.
19	-	% * Compute directional gains on paths from the angles
20	-	% * Implement multi-path fading channels
21	-	% * Perform simple time-frequency channel sounding
22	-	%
23	-	% Submission: Complete all the sections marked \|TODO\|, and run the cells
24	-	% to make sure your scipt is working. When you are satisfied with the
25	-	% results, <https://www.mathworks.com/help/matlab/matlab_prog/publishing-matlab-code.html
26	-	% publish your code> to generate an html file. Print the html file to
27	-	% PDF and submit the PDF.
28	-
29	-	%% Loading the 3GPP NR channel model
30	-	% In this lab, we will simulate a widely-used channel model from 3GPP,
31	-	% the organization that developed the 4G and 5G standards. Specifically,
32	-	% we will use the 5G New Radio cluster delay line model. In the CDL
33	-	% model, the channel is described by a set of path clusters. Each path
34	-	% cluster has various parameters such as an average gain, delay and angles
35	-	% of arrival and departure. The parameters for this model can be loaded
36	-	% with the following commands that are part of the 5G Toolbox.
37	-	fc = 28e9; % carrier in Hz
38	-	dlySpread = 50e-9; % delay spread in seconds
39	-	chan = nrCDLChannel('DelayProfile','CDL-C',...
40	-	'DelaySpread',dlySpread, 'CarrierFrequency', fc, ...
41	-	'NormalizePathGains', true);
42	-	chaninfo = info(chan);
43	-
44	-	%%
45	-	% After running the above commands, you will see that the chaninfo
46	-	% data structure has various vectors representing the paramters for each
47	-	% path cluster.
48	-
49	-	% TODO: Extract the parameters from chaninfo:
50	-	% gain = average path gain in dB
51	-	% aoaAz = azimuth angle of arrival
52	-	% aoaEl = elevation angle of arrival = 90 - ZoA
53	-	% aodAz = azimuth angle of departure
54	-	% aodEl = elevation angle of departure = 90 - ZoA
55	-	% dly = delay of each path
56	-
57	-	% TODO: Compute and print npath = number of paths
58	-
59	-	% TODO: Use the stem() command to plot the gain vs. delay.
60	-	% Each stem in this plot would represent one multi-path component.
61	-	% Set 'BaseValue' to -40 so that the stems are easier to see.
62	-	% Label the delay in ns.
63	-
64	-
65	-	%% Patch Element
66	-	% In this simulation, we will assume the TX and RX patch microstrip
67	-	% antennas. We use the code below to create the antenna element from the
68	-	% antenna demo.
69	-
70	-	% Constants
71	-	vp = physconst('lightspeed'); % speed of light
72	-	lambda = vp/fc; % wavelength
73	-
74	-	% Create a patch element
75	-	len = 0.49*lambda;
76	-	groundPlaneLen = lambda;
77	-	ant = patchMicrostrip(...
78	-	'Length', len, 'Width', 1.5*len, ...
79	-	'GroundPlaneLength', groundPlaneLen, ...
80	-	'GroundPlaneWidth', groundPlaneLen, ...
81	-	'Height', 0.01*lambda, ...
82	-	'FeedOffset', [0.25*len 0]);
83	-
84	-	% Tilt the element so that the maximum energy is in the x-axis
85	-	ant.Tilt = 90;
86	-	ant.TiltAxis = [0 1 0];
87	-
88	-	%% Create UE and gNB antennas
89	-	% We will simulate a channel from a base station cell to a mobile device.
90	-	% In 5G terminology, the base station cell is called the gNB and the mobile
91	-	% device is called the UE (don't ask!). We first create a model for the
92	-	% antennas on each device. In reality, both would have an array of
93	-	% elements, but we will just assume one element for each now.
94	-	%
95	-	% To organize the code better, we have created a class \|ElemWithAxes\| to
96	-	% represent the antenna element. This class is basically a wrapper for
97	-	% the AntennaElement class to include a frame of reference and methods
98	-	% to compute gains relative to this frame of reference.
99	-
100	-	% TODO: Complete the code for the constructor in the ElemWithAxes class
101	-
102	-	% TODO: Create two instances, elemUE and elemgNB, of the ElemWithAxes
103	-	% class representing the elements at the UE and gNB.
104	-	% elemUE = ElemWithAxes(...);
105	-	% elemgNB = ElemWithAxes(...);
106	-
107	-	%% Rotate the UE and gNB antennas
108	-	% The response of the channel will depend on the orientation of the antenna
109	-	% elements. To make this simple, we will assume the UE and gNB elements
110	-	% are aligned to the strongest path. Note that when modifying a class
111	-	% you will need to re-run the constructor.
112	-
113	-	% TODO: Complete the code in the alignAxes() method of ElemWithAxes.
114	-
115	-	% TODO: Find the index of the path with the maximum gain.
116	-
117	-	% TODO: Call the elemUE.alignAxes() methods to align the UE antenna
118	-	% to the angle of arrival corresponding to the strongest path.
119	-
120	-	% TODO: Call the elemgNB.alignAxes() methods to align the gNB antenna
121	-	% to the angle of departure corresponding to the strongest path.
122	-
123	-
124	-	%% Get the directivity along the paths
125	-	% We next compute the directional gains along to the paths.
126	-	% The ElemWithAxes class is derived from a
127	-	%<https://www.mathworks.com/help/matlab/system-objects.html MATLAB system
128	-	% object>, which is MATLAB's base class for objects that can handle dynamic
129	-	% data. In constructing link-layer simulations, it is useful to build your
130	-	% classes as system objects. The key method in a system object is the
131	-	% step() method that is called in each chunk of data. In the derived
132	-	% class, you define the stepImpl() method which is in turn called in the
133	-	% step method. For the ElemWithAxes class, we will define the step method
134	-	% to take angles and return the directivity in dBi.
135	-
136	-	% TODO: Complete the code in the setupImpl() and
137	-	% stepImpl() method of ElemWithAxes.
138	-
139	-	% TODO: Call the elemUE.step() method with the angles of arrivals of the
140	-	% paths to get the directivity of the paths on the UE antenna.
141	-	% dirUE = elemUE.step(...);
142	-
143	-	% TODO: Call the elemgNB.step() method with the angles of departures of the
144	-	% paths to get the directivity of the paths on the gNB antenna.
145	-	% dirgNB = elemUE.step(...);
146	-
147	-	% TODO: Compute, gainDir, the vector of gains + UE and gNB directivity.
148	-	% gainDir = ...
149	-
150	-	% TODO: Use the stem plot as before to plot both the original gain and
151	-	% gainDir, the gain with directivity. Add a legend and label the axes.
152	-	% You will see that, with directivity, many of the paths are highly
153	-	% attenuated and a few are amplified.
154	-
155	-	%% Compute the Doppler
156	-	% We next compute the Doppler for each path.
157	-
158	-	% TODO: Complete the doppler method in the ElemWithAxes class
159	-
160	-	% TODO: Use the elemUE.set() method to set the mobile velocity to 100 km/h
161	-	% in the y-direction. Remember to convert from km/h to m/s.
162	-	vkmh = 100;
163	-
164	-	% TODO: Call the elemUE.doppler() method to find the doppler shifts of all
165	-	% the paths based on the angle of arrivals
166	-
167	-
168	-	%% Transmitting a channel sounding signal
169	-	% The code is based on the lab in digital communications. As described
170	-	% there, the TX simply repeated transmits a signal of length nfft. Each
171	-	% repetition is called a frame. We will use the following parameters.
172	-	fsamp = 4120e31024; % sample rate in Hz
173	-	nfft = 1024; % number of samples per frame = FFT window
174	-	nframe = 512; % number of frames
175	-
176	-
177	-	%%
178	-	% In frequency-domain channel sounding we create the TX samples in
179	-	% frequency domain.
180	-
181	-	% TODO: Use the qammod function to create nfft random QPSK symbols.
182	-	% Store the results in x0Fd.
183	-
184	-	% TODO: Take the IFFT of the signal representing the time-domain samples.
185	-	% Store in x0.
186	-
187	-	% TODO: Repeat the data x0 nframe times to create a vector x of length
188	-	% nframe*nfft x 1.
189	-
190	-	%% Create a multi-path channel object
191	-	% To simulate the multi-path channel, we have started the creation of a
192	-	% class, \|SISOMPChan\|. The constructor of the object is already written in
193	-	% a way that you can call construct the channel with parameters with the
194	-	% syntax:
195	-	% chan = SISOMPChan('Prop1', Val1, 'Prop2', val2, ...);
196	-
197	-	% TODO: Use this syntax to construct a SISOMPChan object with the sample
198	-	% rate, path delays, path Doppler and directional gains for the channel
199	-
200	-	%% Implementing the channel
201	-	% The SIMOMPChan object derives from the matlab.System class and should
202	-	% implement:
203	-	% * setupImpl(): Called before the first step after the object is
204	-	% constructured
205	-	% * resetImpl(): Called when a simulation starts
206	-	% * releaseImpl(): Called on the first step after a reset() or release()
207	-	% * stepImpl(): Called on each step
208	-
209	-	% TODO: Complete the implementations of each of these meth
210	-
211	-	% TODO: Run the data through the step.
212	-
213	-	% TODO: Add noise 20 dB below the y
214	-
215	-
216	-	%% Estimating the channel in frequency domain
217	-	% We will now perform a simple channel estimate in frequency-domain
218	-
219	-	% TODO: Reshape ynoisy into a nfft x nframes matrix and take the FFT of
220	-	% each column. Store the results in yfd.
221	-
222	-
223	-	% TODO: Estimate the frequency domain channel by dividing each frame of
224	-	% yfd by the transmitted frequency domain symbols x0Fd. Store the results
225	-	% in hestFd
226	-
227	-	% TODO: Plot the estimated channel magnitude in dB. Label the axes in
228	-	% time and frequency
229	-
230	-	%% Estimating the channel in time-domain
231	-	% We next estimate the channel in time-domain
232	-
233	-	% TODO: Take the IFFT across the columns and store the results in a
234	-	% matrix hest
235	-
236	-	% TODO: Plot the magnitude of the samples of the impulse response
237	-	% in one of the symbols. You should see a few of the paths clearly.
238	-	% Label the axes in delay in ns.
239	-
240	-
241	-	%% Bonus: Viewing the channel in delay Doppler space
242	-	% Finally, we can estimate the channel in the delay-Doppler space.
243	-	% This is commonly done in radar and we will use the exactly same procedure
244	-	% here. In the frequency domain response, hestFd, each path results in:
245	-	% * Linear phase rotation across frequency due to delay of the path
246	-	% * Linear phase rotation across time due to Doppler
247	-	% Hence, we can see the paths in the delay-Doppler space with a 2D IFFT.
248	-
249	-	% TODO: Take a 2D ifft of hestFd and store in a matrix G.
250	-
251	-	% We can now extract the delay-Doppler components from G.
252	-	% Most of the interesting components in a small area:
253	-	%
254	-	% nrow = 64;
255	-	% ncol = 32;
256	-	% Gs = [G(1:nrow,nframe-ncol:nframe) G(1:nrow,1:ncol)];
257	-	%
258	-	% TODO: Plot the magnitude squared of Gs in dB using imagesc. Label the delay
259	-	% and doppler axes. You should see the components clearly.
260	-
261	-	% We can even compare the peaks in the matrix Gs with the delay and Doppler
262	-	% of the actual components.
263	-	% TODO: Find the indices of the paths with top 20 directional gains.
264	-
265	-	% TODO: On the same plot as above, plot a circle corresponding to the
266	-	% (doppler,delay) for each of the top components. You may have to reverse
267	-	% the doppler due to the sign conventions we have used.
268	-

unit03_fading/partial/lab_chan_sounder.mlx

Binary file.
unit03_fading/partial/lab_chan_sounder.pdf

Binary file.

Changed m to mlx file