Month: May 2017

At OneTwo, we are always searching for ways to make things easier on the workfloor.  For nearly a decade, we introduced barcode scanners everywhere! From architects and accountants, over carpenters and contractors, to farmers and gardeners are now using our small barcode scanners to record times, jobs and used materials.

The last couple of weeks, I’ve been investigating how we could tackle another common issue, but now on the factory shop floor.  Working on a specific step in the development of a product demands the constant focus of the operator so using a barcode scanner to get the instructions is, well, just not practical if you have your hands full.

Time to get to the lab! Provided with heaps of coffee, some unhealthy snacks and an open mind we started putting some ideas together on the whiteboard.  We wanted to give as much feedback to the operator as possible using monitors,  beamer projections at the workstation and personal instructions on phones and tablets.

The key to deliver hyper-contextual content to the users is knowing the location of every chess piece in the game: the workstation, the machine that is going to be build and the operator(s).  The underlying communication technology was going to be Bluetooth Low Energy (BLE). The Raspberry Pi has it and so does almost any phone or tablet. By attaching a cheap BLE Beacon to the machine we could understand the location of all the players.

Using B4X for this project was a no-brainer, as Erels toolbox is just made to build stuff like this! One important link in the chain was missing:  the BLE reader in B4J for the Raspberry Pi.  However, writing some small scripts and calling it using jShell and a couple of timers did the job just fine.

Scripts code:

Two scripts to discover BLE devices around the Raspberry Pi.  Each one runs in its own jShell in B4J. First one scans for devices, second one reads whatever they broadcast.

#!/bin/bash
# Beacon Scan by Alain Bailleul 2017
sudo timeout -s SIGINT 5s hcitool lescan

#!/bin/bash
# Beacon Scan by Alain Bailleul 2017
sudo timeout -s SIGINT 5s hcidump -X > scan.txt

Also, as the Raspberry Pi has to be a BLE beacon itself (so the Android Native app can discover it), I had to write another little script:

#!/bin/bash
# Beacon Scan by Alain Bailleul 2017

sudo hciconfig hci0 leadv 3
sudo hcitool -i hci0 cmd 0x08 0x0008 19 02 01 06 03 03 aa fe 11 16 aa fe 10 00 02 6f 6e 65 2d 74 77 6f 07 $1 $2 $3 00 00 00 00 00 00
sudo hciconfig

IPN=$(ip addr show eth0 | awk ‘/inet / {print $2}’ | cut -d/ -f 1)
IPW=$(ip addr show wlan0 | awk ‘/inet / {print $2}’ | cut -d/ -f 1)
MACN=$(ip link show eth0 | awk ‘/ether/ {print $2}’)
MACW=$(ip link show wlan0 | awk ‘/ether/ {print $2}’)

echo “$MACW;$IPW;$MACN;$IPN;END” > macip.txt

For anything related to presenting the instructions, we could use ABMaterial.

Alright! We’ve got everything we needed to set up a test scenario for a workstation so time to bring out the cool slide:

Although there are a lot of physical components and different OS’s in play, B4X has all the tools to make them seamlessly talk to each other using one language.  For the native Android part, I wrote my own BLE discovery library based on the altbeacons library. (Note, there is a library available in B4A to discover BLE devices).

For the video demonstration, things you have to keep in mind:

INPUT:

1. OneTwo box + BLE beacon: represents a machine arriving on the workstation
2. Raspberry Pi (under the table): represents the workstation
3. Android Phone: represents an operator/visitor at the workstation

OUTPUT:

1. On the monitor, for each ‘machine’ some info is presented (e.g. a serial number, to what country it has to be shipped, etc)
2. On the phone: specific instructions for the operator(s) for this ‘machine’, or a greeting to a visitor.
3. On the workstation, projected global instructions for the operator(s) for this ‘machine’ with a beamer.

Let’s have a look how all of this plays out:

This turned out to work all very well for a first trial! For real world usage, the project will need some more work (like improving the algorithms to discover the BLE devices, setting thresholds or calculating them without calibration, etc…).  But we are very excited with the possibilities this low-cost solution can already show in such a short time.

Next step will be setting this up on a real workstation on the shop floor in a real factory later this month.

Until next time,

Alwaysbusy

Click here to and get the latest ABMaterial first!