I can't recall offhand if it came with CircuitPython or micropython (apologies, it has been a bit since I worked with the car). If it was the former I think you'll need to modify the script to work with the other or install micropython first to run it. IIRC I installed micropython myself and then ran the logic.

Hey they work great. There's the issue with partial refreshing but with the full refresh it has been pretty consistent. There's also a PR someone mentions in this thread re: the 3 color displays. I have a few of those so hoping to try it out in the future. github.com/esphome/feature-requests/issues/2101

@thecosmicbee definitely looking forward to seeing how those work! I've been told their refresh is a bit slower but for most data statistic displays that shouldn't be an issue unless if it's like over 5minutes or something absurd like that. Either way, great work and I'll sub for future videos! Best of luck to you and your projects!

Yeah, taking a photo of the sand each time on change for use with DepthAnythingV2 with ControlNet using the automatic1111 standard diffusion web UI with API to handle the rendering. Detecting hands in the photo each iteration to determine if it should rerender (once hands leave the viewport).

Nevermind, I see now. You can directly input what it should be. Very cool

Using DepthAnythingV2 -- it takes a photo and uses opencv to warp it so it's constrained to the sandbox based on the CharucoBoard projection. Then with that it uses it as the input to the standard diffusion ControlNet Depth for a set prompt. The prompt can be updated via voice.

Thank you!

Yeah for sure, in retrospect I should have used one of the custom commands and taught it to just "Stop" over that built in "Park a Car" which feels a bit harder to parse. Definitely had me sweating when it ignored me the first time haha. I think the wheels on the floor from the mecanum car make it a bit harder to hear as they were pretty loud. I think if I were to do it again I'd make the voice part a handheld device that then just sends the commands to move over its hosted access point network perhaps with custom commands for shorter movement like advancing only a few feet in that direction each time.

​@@thecosmicbeeI like that idea, just have a headset on and send the commands. Then connect multiple robots together in Ardupilot Rover and have a voice controlled robot swarm.

They have a python and arduino library for the sensor. github.com/DFRobot/DFRobot_DF2301Q my rust module is located at github.com/Cosmic-Bee/df2301q.rs docs.esp-rs.org/book/installation/rust.html has some information about getting setup for developing on the esp32 platform for rust (I used it here for this module). The arduino library may be the easiest path forward though as you can do a search from the arduino ide to install that and get up and running fairly quickly.

Hey there, sorry for the late reply, right now it's in a fork of the SenseCAP Indicator ESP32 repository as a pull request. I have the pull request: github.com/Seeed-Solution/SenseCAP_Indicator_ESP32/pull/24 The associated branch: github.com/Timo614/SenseCAP_Indicator_ESP32/tree/timo614-indicator-matter I recently updated it to a more recent esp idf version as the underlying repository was updated as well.

I had originally seen a similar setup from the Ruiz Brother's darksaber tutorial so knew I could handle LED strips from the propmaker itself and went with a similar setup here: learn.adafruit.com/ble-darksaber-propmaker/circuit-diagram I just took a look at the schematic now to see if perhaps it had some boost element but it doesn't look like it does: learn.adafruit.com/assets/69243 It looks like it sends either the vbat or the vbus to the v+ line which is then used for the neopixel out port. So yeah nothing there boosting it. I suppose it must be powering it with whatever voltage the battery has and it must be enough for the ws2812b strips (I had thought they were 5V myself and it was boosting it).

Hope to see more!

Hey, this was a LED strip for a poolside lane indicator for night swimming. Hackster had a contest for creating some devices for assisted living and one of the topics was visual issues while swimming. Went super low tech with this one as I ran out of time. I had originally hoped to do something with machine learning to help indicate the user's location in the lane and then update the LED strip (so perhaps something like flooding the entire right side red to indicate you've gone out of the lane).

Yup, from talking with Vivek one of the big issues he had was with normal mice and the movement involved for both positioning and clicking. The conductive paint worked well for translating the touch elements making for a mouse that triggers easily and responds well with little movement on the user's part.

You can configure the maximum duration for the active state with the set wake time command. From their header file it looks like 255 seconds may be the limit: github.com/DFRobot/DFRobot_DF2301Q/blob/master/DFRobot_DF2301Q.h#L168 You can setup a new wake word and it'll learn it. I haven't tried clapping for the wake word but I think it would work. You can't configure the opening phrase it says when it boots but for low key setups it seems like a really neat way to get up and running for simple voice commands without needing to do any ML. I wish there was a way to disable all of the commands and then enable a subset as it's a bit weird for it to respond to things like "Park a Car" when I'm using it as a trigger for lights in my kid's room.