Repository navigation
Replies: 1 comment 1 reply
|
Hi @VS511, Sionna RT can compute gradients with respect to the transmitter or receiver position (see example below), however these will be incorrect because discontinuities due to change of paths visibility are not considered (paths that become occluded or unclouded). scene = load_scene(rt.scene.simple_street_canyon)
scene.tx_array = PlanarArray(num_rows=1,
num_cols=1,
vertical_spacing=0.5,
horizontal_spacing=0.5,
pattern= 'iso',
polarization="V")
scene.rx_array = scene.tx_array
tx = Transmitter(name='tx',
position=[45,0,1.5])
scene.add(tx)
rx = Receiver(name='rx',
position=[-40,0,1.5])
scene.add(rx)
solver = PathSolver()
frequencies = subcarrier_frequencies(1024, 30e3)
solver.loop_mode = "evaluated"
dr.enable_grad(rx.position)
paths = solver(scene)
cfr = paths.cfr(frequencies)
e = dr.sum(dr.square(dr.abs(cfr[0])) + dr.square(dr.abs(cfr[1])))
dr.backward(e)
print(rx.position.grad)Output: |
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Are receiver positions differentiable in Sionna? AFAIK, they’re not. I'm trying to estimate the location of a CFR measurement with unknown Rx position (as is the case in cellular scenarios). Any ideas on how to approach this?
Thank you!
All reactions