First OTA Run

Before the First OTA Run

Prepare two backend nodes. In downlink-only mode, BS uses one TX plus one sensing RX antenna path and UE uses one RX path. If duplex/uplink is enabled, UE also needs a TX antenna/RF chain and BS needs an RX antenna/RF chain for the uplink; in FDD mode the radios must also support the configured uplink carrier.
Start from the sample YAML that is closest to your hardware. The repository includes X310/B210 examples, but it is not limited to those USRP models.
Copy runtime YAMLs into build/, because both binaries read BS.yaml or UE.yaml from their working directory.
If the frontend runs on another machine, point the Python viewer at the backend with --host <backend-ip> or the viewer’s Backend IP field. default_out_ip is only for destination-style output IPs, not UDP/ZMQ listen addresses.
If you care about stable real-time behavior, run scripts/set_performance.bash, then apply CPU isolation with sudo ../scripts/isolate_cpus.py (asks BS/UE/BS+UE, or pass --role), and only then launch via sudo ../scripts/isolate_cpus.py run ....

Typical Usage Example

1. Startup of the BS

sudo -s
cd build
# For X310:
cp ../config/BS_X310.yaml BS.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./BS

# For B210:
cp ../config/BS_B210.yaml BS.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./BS

# For B210 duplex:
cp ../config/BS_B210_Duplex.yaml BS.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./BS

If you are using a separate frontend computer, point the monostatic viewer at the BS backend IP with --host or the viewer’s Backend IP field.

2. Startup of the UE

sudo -s
cd build
# For X310:
cp ../config/UE_X310.yaml UE.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./UE

# For B210:
cp ../config/UE_B210.yaml UE.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./UE

# For B210 duplex:
cp ../config/UE_B210_Duplex.yaml UE.yaml
sudo ../scripts/isolate_cpus.py
sudo ../scripts/isolate_cpus.py run ./UE

If you are using a separate frontend computer, point the bistatic viewer at the UE backend IP with --host or the viewer’s Backend IP field. Set default_out_ip only for decoded/debug outputs that should be sent to another machine.

3. Stream video to the BS

ffmpeg -re -stream_loop -1 -fflags +genpts -i video.mp4 -an -c:v libx264 -x264-params keyint=5:min-keyint=1 -b:v 3000k -minrate 3000k -maxrate 3000k -bufsize 1M -f rtp -sdp_file video.sdp "rtp://<your IP of the BS>:50000"

If you are streaming locally, your IP of the BS can be set to 127.0.0.1.

If you want to stream audio together with the video, use RTP MPEG-TS instead:

ffmpeg -re -stream_loop -1 -fflags +genpts -i video.mp4 -c:v libx264 -x264-params keyint=5:min-keyint=1 -b:v 30000k -minrate 30000k -maxrate 30000k -bufsize 1M -c:a aac -b:a 128k -ar 48000 -ac 2 -f rtp_mpegts "rtp://<your IP of the BS>:50000"

4. Play video from the UE

Copy video.sdp to the video receiver, modify m=video 50000 RTP/AVP 96 to m=video 50001 RTP/AVP 96.

ffplay -protocol_whitelist file,rtp,udp -i video1.sdp

Note that this command should be run on the frontend.

For the RTP MPEG-TS stream with audio, you can play it directly without an SDP file:

ffplay rtp://0.0.0.0:50001

5. Run monostatic frontend

python3 ./scripts/plot_sensing_fast.py

This maintained viewer auto-selects CUDA, MLX, Intel GPU, or CPU backends.

6. Run bistatic frontend

python3 ./scripts/plot_bi_sensing_fast.py

This maintained viewer auto-selects CUDA, MLX, Intel GPU, or CPU backends.