Group Actions
Engagement
  • Author Author: bernhardmayer
  • Date Created: Date Created
  • Views 568 views
  • Likes 6 likes
  • Comments 4 comments
Related
Recommended

PlantProtector3100 - Bernhard - Bluetooth Unleashed #1 - Introduction

bernhardmayer

Description of the Problem

Plants need water to develop well and stay healthy.

 

Healthy and not so healthy plant:

image

 

This sounds very simple and everyone understands it. Neverless plants sometimes have to suffer dryness simply because the owner has forgotten them. Or they get to much water so that they drown.

 

In a private household watering the plants is a managable task but also creates some trouble. When you go to larger buildings like office spaces or public buildings, plants are always forgotten and have to die simply because of the sheer amount of plants.

 

PlantProtector3100

This is where my PlantProtector3100 comes in.

 

System overview:

image

The PlantProtector has to be put into the soil in the pot of every plant and measures the humidity and temperature of the soil. The devices are connected via Bluetooth. To cover larger areas I want to create a Bluetooth mesh network where each device works as a range extender for the next device. So my sensors will also work in the larger buildings like office spaces. At one point of the network it is connected to the internet and will deliver the data of the sensor to a cloud service. The user can connect to the cloud service and see all the data of the sensor and get alerts if data is out of a specific range like when the soil gets to dry and watering is necessary.

 

Hardware setup with STSW-BNRG-Mesh

To get a bluetooth mesh network running you some mesh network library or SDK. At the moment it seems like ST is the only company with an SDK so I will give it a try. They call it

STSW-BNRG-Mesh. You can find more information on it here: STSW-BNRG-Mesh - Mesh over Bluetooth Low Energy - STMicroelectronics

As development board i decided to use the STM -  NUCLEO-L073RZ and X-NUCLEO-idb05a1 I have addional evaluation boards and hardware from ST available that should support STSW-BNRG-Mesh and i will introduce them in my next post As internet gateway I will use a Raspberry Pi I have to figure out if I can use the built in Bluetooth hardware of the Raspberry Pi to connect to the mesh network or if I have to use an additional ST SPBTLE-RF at the Raspberry Pi to connect

 

Design goals

  • low energy consumption of the sensor nodes so that the battery will last long, hopefully for the whole lifetime of the sensor
  • high range of coverage of the bluetooth network
  • high amount of supported single sensor nodes
  • reasonable detection of moisture

 

Project plan

  • getting the evaluation boards to work
  • setting up the mesh network
  • connnect the gateway to the mesh network
  • set up a cloud service like Google Cloud or Amazon Web Services and connect the data of the mesh network
  • develop a reasonable moisture sensor
  • make own pcb / device with everything onboard (smaller than the evaluation board and adapted to the environmental conditions at the plants)

 

Challenges

If the mesh network doesn't work like show in the demo and expected my project is at risk. I could switch to a standard bluetooth network setup where each single sensor node is connected to the master/gateway (the Raspberry Pi) . This would also work somehow but the mesh network has more charm and technical advantages.

  • You could also have a look at OpenHAB as a control platform. It could read in all the data from 100’s of sensors; graph them out, and provide cloud connectivity. You can run the server on a RaspberryPi.

    on top of that, it offers a pretty robust and powerful rules engine. You could, say, turn on a light at each unit requiring water.

  • The Zwave stuff i’ve worked with only supported “mesh” / “repeater” modes if they were wall-powered. The battery devices didn’t have it. At least not on the few things I’ve used. Obviously acting as a repeater could use more power than just “wake up and report status” then going back to sleep.

  • Nice start.

     

    Have you considered doing NDVI imagining to verify that your plant is under stress to help calibrate your stress analysis?

     

    DAB

  • Nice start to your design challenge.  Mesh networks offer a lot of advantages, when it comes to large and convoluted networks.  But they can also present some interesting problems/issues.  I worked on a Zigbee mesh solution to office/parking garage/warehouse lighting systems. The mesh was very good at finding good connectivity through some of the complex floor plans that we installed our systems in, but sometimes there could be throughput issues as message 'storms' could pop-up, were message throughput would suffer.  Some of the solutions required 'throttling' message rates from individual devices such that they did not over stress the network.  Hopefully with your smaller, less dense network, these will not be problems for you.  It is nice that you have already begun to consider work arounds should the mesh not work for you.

     

    Good luck on you project!

    Gene