So I have the bot assembled, but there are some software issues I’m not sure how to approach. Specifically I never get any output to the OLED, and the Jupyter instance only seems to start if the bot is connected to an HDMI display. I did have to install the Python serial package, and that dealt with one error message I saw when I ran stats.py manually, but it bails out quickly because it can’t read the wlan1 info. Not sure if I should just disable that part of the script or if the absence of the file it’s trying to read is indicative of something more serious.
I walked partway through the basic motion demo and things seemed to be working. The object-following demo complained about missing files, but I didn’t delve too deeply into that yet
If the script is failing because it can’t read wlan1 info, it could be because wlan1 is not available on your system or it’s named differently. Check the network interfaces on your Raspberry Pi using the ifconfig or ip addr command to see if wlan1 is present. If it’s not, you may need to modify the script to use the correct interface name or disable that part of the script if it’s not essential for your bot’s functionality.
I went through that script and the utilities it uses last night and it looks like that message is informational rather than critical. From the behavior I’m seeing it looks like the Jupyter service and the stats service are both not launching when the Jetson boots headless, which is disappointing since in theory all that was supposed to be necessary was to configure the WiFi. That may have to wait for the weekend. Tonight my plan is to try to figure out why even when running with a display the stats.py script doesn’t emit anything to the OLED. Random guess is that it’s using the wrong I2C address. Time permitting I also need to double check connections to see why one of the motors isn’t live.
Looks like some of my software issues are hardware issues. The kit appears to have been delivered with a non-functional OLED breakout and one of the two motors dead. I’m not impressed.
Incorrect screws.
Broken components.
Incomplete SD image.
It appears I’m supposed to cut a trace on either the OLED board or the motor controller based on the hookup guides for those components but not mentioned in the kit instructions.
Anyone have any advice on getting the jupyter service to run without a monitor connected?
So here’s a fun follow-up. I got a new OLED module to replace the dead one from the kit. Plugged it in this morning and turned the bot on just for the heck of it. The good news is that the new OLED works. The unexpected news is that the Jupyter server starts up now without being connected to a monitor.
That’s a side-effect of the headless mode…if a monitor isn’t hooked up, the jupyter server should launch by default. If that isn’t desired you could always disable it as well
That’s the part that’s weird, though. The Jupyter server autolaunched fine with a monitor connected but didn’t launch headless until I attached the working OLED breakout.