Support for pico w wifi

[Manos-G]

Hi @hippo5329,

I’ve been exploring the possibility of using the Pico W with Wi-Fi transport and noticed in the commit history (specifically commit 32f972abba8ef31b1154c0e5f3dd760c67268e57) that you were working on something similar.

Did this approach face any specific issues or limitations? Any insights you could share would be greatly appreciated

Thanks in advance!

[hippo5329]

I had issues with the wifi driver on pico-w. I would not suggest pico-w. The wifi on pico-w is an SPI add-on. So the bandwidth is very limited. It won't be useful for real micro-ROS jobs. I would suggest esp32 if you need wifi.

[Manos-G]

Thanks for the insight! That’s helpful to know.

Do you think the bandwidth limitations would still pose a significant issue if I’m not using something bandwidth-intensive like a LiDAR?

[hippo5329]

The linorobot2_hardware control loop timer runs at 50Hz. The latency is critical. We have to increase the serial baudrate to 921600 for serial transport.

The typical high clock rate of SPI is 1Mhz. The achieved data rate might be less than 100K bits/sec. It might be fine for serial print or simple int32 publisher at very slow rate. But it is way too slow for robot control loop.

The esp32 wifi can run above 10+ M bits/sec in most cases. It is simply not worth the work to use pico-w.

[Manos-G]

Thanks for explaining that so clearly! Makes total sense now why the ESP32 is a better fit. Really appreciate you taking the time to share this—it’s a huge help!