Meshposition framework: starting at the very beginning

Hello! I just found this group. What a great wealth of knowledge!

I have what I hope is a straightforward sport application: Track one tag at ~5hz within a 60m x 40m rectangle. I’ll be using range and direction information to control my stepper motor pan head.

I’ve read through the all of the #networks:ultrawideband posts and the HomeSmartMesh site and haven’t found the answers to a few basic questions:

  • Which Qorvo dev systems to buy in 12/2021 and how many nodes?
  • Can a single RasPi running meshposition serve as a standalone Gateway/Master Orchestrator for the entire system?
  • I’d like to have that RasPi also handle my hardware control (stepper motor logic, switch inputs, LED status lights) but that could be offloaded to a dedicated esp32-type controller if necessary.
  • Is meshposition in a stable state or is it still in development?

For reference, I’m “somewhat technical”: I can build a tool chain, flash a microcontroller boot rom, etc. Preferred deployment environment is Python/Microphython with C-language optimizations for speed critical functions. Strong background in 3D graphics. Not a fan of Arduino. Developed apps in Swift for iPhone. Zero knowledge of Android ecosystem.

Thank you and I apologize in advance if this information is covered elsewhere.

Hi @tritip ,
thank you for your interest in the content of this forum and website.

I’ll try to answer your questions but I’m afraid that you could be a bit disappointed from the fact that meshposition does not even qualify as being still in development as it is not even intended to become a production software. It is at best a proof of concept and a hobby or academic research platform.
Now given the open source nature and the licensing type, anyone of course can make a commercial product out of it and I will even help, but my contribution is limited as I do not intend to spin this project to a lucrative income source. The reason I started this project is that I found no open and flexible platform so wanted to prove to the community that it is possible, but that does not compete with existing commercial solutions.
If you explain me your context and business model, I can also give advices and recommendations. For example, if you’re a startup you’d need different tools than if you’re subcontracted from a big company. Also if you’re a student then you might have a pedagogical goal rather.

  • Your first question about current Qorvo dev systems will be highly welcome in the official forum I think as their company employs will answer you and give you free guidance.
    I have created a survey of UWB dev kits here Ultra Wide Band | Home Smart Mesh but did not update it the last two months.

  • second if a single RasPi can do it, yes. The idea is that the python layer only creates a request which will be executed by every node through a time in real time, so the RasPi does not have any real time constraints when it comes to the ranging data collection protocol, it is performed by the nRF nodes.

  • So as per the previous answer, here also yes, the raspi should have enough power to handle any other control.

  • last question is what I answered firs in the intro, unfortunately no, meshposition is not even intended to become stable. The only serious issue that’s missing is an acknowledged level2 QoS RF mesh protocol. Currently all nodes have to be within reach which does not achieve yet the main target of the concept of propagating commands through mesh. Although this is not hard to achieve, I hesitated a lot between waiting for a 802.15.14 enabled chips or start a custom RF with the nRF52832 that I have that are not 802.15.4 capable.

Other than that, if you’re interested in general in such types of project you see on the website or you have ideas you’d like to brainstorm about, feel free to post and we can discuss about them.

1 Like

Thank you for taking the time to respond in detail.

I’m an hobbyist programmer working at home. My sports application is designed to help my wife document her sports training.

Based on my own reading of the Decawave forum I’m leaning towards buying a MDEK1001 kit. It’s not the newest technology but it’s available in January and not terribly expensive.

It seems like I’ll be able to configure their existing software for a network of multiple anchors that will broadcast the location of a single moving tag over MQTT. If it’s reasonably accurate then I’ll have everything I need for my little project.

Thanks again for creating this forum. It’s been surprisingly difficult to find any information on this subject.

I’m glad these info helped you. MDEK1001 is exactly the one I have the with 12 units.
I totally recommend the default firmware and the default demo app. I was quite happy with the experience, I got everything up and running in a moment and with the raspi gateway, I could get mqtt updates. Now pay attention of the raspi 3 model you need to use, a newer with 1gb eth will not work. And the firmware is not flexible. You might wonder sometimes why the reported tag position is what it is, it has wierd filtering. But if your requirements are not high, you will not have to deal with a custom firmware.
By the way, I’m doing these projects as a hobby too, so I like to discuss similar hobby projects. Especially if you take yours to the level of sharing it somewhere e.g. github, hackaday, maybe I could even contribute.