So I built myself the PiDog Raspberry Pi based and recently the GalaxyRVR Arduino based.
Both working like a charm and enjoyed that quite a lot.
Learned a lot on the mechanical and controlling part of the motors, sensors etc.
I also have been writing a python based voice controlled AI assistent and that went quite well as well. I did this on the Pi5 which I mounted in the Pironman5.
With the GalaxyRVR it is built to be controlled by the app which is sending (I think REST type) commands via Wifi to the ESP32 that is preloaded with SunFounder library to actually fit the rover.
What I like to do now though is to place some commands inside my written voice controlled assistent to control the rover.
Can someone put me on track how I could be doing this?
Do I need to flash the ESP32 with own or standard library?
Can I add extra sensors somehow on the RVR hat or should that be done directly on the Arduino making use of the unsoldered pins?
Not sure that I’ve totally understood your objective, for example I have no idea what a GalaxyRVR is! but assuming that I have!
I added voice control to the pidog, see here
I used a usb microphone and the vosk voice recognition database, as it’s a very small footprint. I used my own edited version of Sunfounder’s keyboard code for the control.
Sunfounder’s chatgpt code also has some good examples on voice control.
It looks like something I would like to achieve as well I guess. Never looked to ROS yet though and don’t think it is required for what I want to do.
I’m indeed already using a small usb microfoon and have a nice local STT and TTS model already up and running even a local LLM that the Pi5 just can handle reasonably as well.
My aim here is to not have everything perfect but to learn the basics and probably should look into ROS as well. Eventually I want to make a complete own design. Mechanically, electronics and coding with all bells and whistles.
Probably will do that with a nVidia Jetson for example. But I’m far far from there still One can dream big right……
The GalaxyRVR is another product from SunFounder by the way
Do you got some GitHub location maybe where I can snif around on the things you did?
Hi, no I dont have a github location. You are 100% correct, ROS is not needed for what I did, in fact it just adds unnecessary complexity. I only did it via ROS as a platform for myself to learn ROS. I am considering rewriting the code outside of ROS such that I can share it with others as standalone.
I do use the inbuilt ROS joynode to publish the joystick commands, and the ROS publish/subscribe topic methods, so I’d just need replacements for those.
However , I’m currently developing VSLAM for my pidog, which is consuming my spare time, so in reality I’ll probably not return to my joystick and voice control.it works fine for my needs, albeit a bit complex.
