(14/7-18/7)
We continued working on using the new Arduino R4 WiFi board to upload our code to Uno WiFi to connect to our network, for the caregiver alert system. We scaled the usual 2h down to 20s for the purposes of testing, meaning we received these alerts when no movement was detected for 20s.
We needed to troubleshoot many times when the sensors went haywire and did not sense weight shift, wires came loose, and vibration motors did not work. By Wednesday (16/7), most of our cloud function was successful and we linked our network to the Arduino WiFi. We found that Arduino IDE and Cloud worked differently.
We also managed to connect the Arduino setup to a portable charger instead of our laptop, which will also be in the electrical box (storing the wires and breadboard) by the side of the bed for neat wire management.
The load sensors were initially (rather poorly) secured to the foam by cutting holder sized holes into the foam, placing pieces of cardboard in the holes to create a level surface, then placing the holders inside. After we finalised that all sets were working and no more wires would come loose, we hot glued the cardboard pieces into the foam holes and the holders onto the cardboard and along the sides of the foam holes.
Hard at work!
OWWWWW Life without our CSO….:(
We scaled up the time factor of our testing (initially 20s) as we were afraid the 20s cycles were too short to account for delays in sensing of weight change (ie movement) which would result in unnecessary vibrations. SEE HOW
Also, currently a change in the reading of one load sensor set of at least 2kg constitutes movement and prevents vibration motor activation. However, this could be inaccurate as it could have been caused by minor non-positional changes. Hence, we worked on increasing the movement threshold to requiring 2kg weight change in at least two load sensor sets to constitute movement. We did this by ensuring if there was weight change of 1kg each on two sets the motors would still vibrate.
We got feedback on the urgency of our wire management, as at this point we were using a cramped breadboard and wires from different groups were entangled. Hence, there were worries about connections coming loose due to strain when we have to move our prototype.
This was the initial state it was in…yikes.
We decided to get rid of our breadboard and connect the DuPont wires, reinforcing them using heat shrink.
Some wires became really long in the process.
After bundling and taping the groups of wires together, this was our new breadboard-less wire setup, which we went on to compact in order to gauge the dimensions for the electrical box. We had to ensure that the wires were not stretched too much where it affected the vibration motors or sensors attached to the bed, to prevent breakage.
The dimensions we obtained were 17cm (length) x 12cm (width) x 12cm (height) but adapted the dimensions in Fusion based on existing designs in the end.