Mimic Platform Project
Shared with express permission of Curtis Berlinguette Research group
Last updated
Shared with express permission of Curtis Berlinguette Research group
Last updated
In my first few months at Ada I noticed that 24/7 operation was severely limiting our hardware development work. I proposed that we build a duplicate hardware platform out of our back-up modules which would allow us to iterate on hardware design without sacrificing system up-time.
The base of the Ada robot consists of a SCARA manipulator arm and heavy duty platform with standardized mounting holes. Multiple units of this system were provided to us by the vendor. My first task was to contact the vendor and conduct hardware setup of our backup unit as well as put together our existing Python control environment on a dedicated control terminal.
After pitching this project one consistent feedback item was a desire for an integrated utilities control system. Previously we had been designing a separate actuation system for each and every Ada hardware module. Some users were even 'borrowing' vacuum or air supply lines from the Ada robot to test their modules during system downtime. It was essential that the mimic platform package these utilities alongside the primary robot manipulation arm. From personal experience I was also aware of how important manual control is for sketching out movement routines. Therefore I set out to make utility supply control panels which could be controlled programmatically and manually.
I created the following instructional videos to supplement a user guide I had written. They demonstrate the functionality of the final iteration of each utility box:
I conducted a risk assessment in order to identify and address the safety challenges posed by this project. I created the table below and spent ~1 hour filling it out (various items omitted for brevity). I then scheduled a meeting with various project stakeholders (co-workers) and collected additional items and comments. I identified the priority hazards to address and implemented mitigations into my design.
The 'first customer' of the mimic platform was this uncapping/capping routine which had previously been in and out of development for over a year due to reliability issues. Using the mimic platform I was able to iterate on the existing routine and clamp design without concern for blocking system uptime. This gave me the freedom to run extended reliability tests. I was able to demonstrate >100 consecutive successful cap/uncap operations which met the established reliability requirements of the group.
...
| Hazard | Impact | Likelihood | Priority |
|---|---|---|---|
A human is pinched by fixture movement
0.80
0.75
0.60
A human is shocked by a motor controller (<60V)
0.30
0.50
0.15
Glassware is shattered by robot movement
0.05
1.00
0.05
A human is shocked by the N9 arm (>100V)
0.90
0.05
0.045
