streampipes-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philipp Zehnder <zehn...@apache.org>
Subject Re: Tools to retrofit machines with StreamPipes
Date Mon, 13 Jan 2020 21:47:01 GMT
Hi,

just in case some of you are interested in gathering power information of machines and add
human generated label to the data.

I found a cool app called MQTT Dash, which we use instead of hardware buttons for user input
next to machines. 
Originally it was designed as a dashboard app for your smart home and is very simple to use.
To gather user input (e.g. scrap parts on a machine) create a dashboard with buttons operators
can press. For each interaction an event is triggered which is read by a StreamPipes Connect
adapter.
Those events are stored in a database and can later be used as labels to train machine learning
models.

For the power sensors, I found those [1]. It seems that they can generate high frequency events
of the current power consumption of the connected machine. There is also a MQTT version available.

Cheers, 
Philipp


[1] https://www.netio-products.com/en/products/all-products <https://www.netio-products.com/en/products/all-products>
  

> On 31. Dec 2019, at 11:46, Philipp Zehnder <zehnder@apache.org> wrote:
> 
> Dear all,
> 
> we already have plug-and-play solutions for several sensors that can be used to retrofit
machines and build fast prototypes with StreamPipes.
> For example the Bosch XDK sensor, the TI Sensortag, or the raspberry pi camera are already
integrated.
> 
> Now I am looking for a plug to measure power consumption (1 Event /s) and a button to
get user input (e.g. to label data).
> I already found some bluetooth or wifi solutions, but it seems all of them require a
custom integration.
> I was wondering if anybody of you already has some experience and could give me a hint
of an easy to integrate solution. 
> Ideally the devices would already support some standards like MQTT, but it would also
work to send the sensor values via bluetooth to a raspberry pi and then from there to Kafka
or MQTT.
> 
> Please let me know if you have any further ideas for sensors that we could integrate.

> 
> Cheers,
> Philipp


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message