Fall Validation Demonstration
Objective: Ensure that the entire system functions as expected and can sufficiently perform the
described use case for an “average user” under normal circumstances.
Procedure:
- During the test an assistant will record successes, failures, measure errors, and time procedures
- User will put on Coborg backpack
- User will hold a matte panel overhead in one corner of the task space and say, “Coborg, move to target”
- Coborg will operate
- User will move 6 inches in cardinal directions
- User will rotate 90 degrees to the left, followed by returning to the center, rotating 20 degrees to the right, and
returning to the center again - User will screw in the panel and say, “Coborg, go home”
- User will place hands on panel and say, “Coborg, move to target”
- Coborg will operate
- User will unscrew panel and say, “Coborg, go home”
- User will flip panel to shiny side and repeat previous steps for a diagonal corner of the task space
- Test will be repeated as time allows
Verification Criteria:
● Panel is successfully screwed into the board all times without interference from the user
● Euclidean distance error to target is less than 6” in ≥ 60% of trials
● Target is received less than 5 seconds after command issue
● End-effector will shift less than 6” during screwing process
● The robot is able to hold the part overhead without dropping it
● The percentage of all commands correctly identified (out of all true/false positives and false negatives) is ≥
60%
● Both the number and length of commands are inherent to the test
● Release of part takes less than 5 seconds
● Point of arm furthest from base motor will be ≤ 20” in compact mode
● User will be unharmed during test, also inherent to system (e-stops, torque limits, etc.)
● User does not have difficulties, automation works as intended
● Robot battery performs for at least 20 minutes or less than a corresponding percentage of the battery has
been used during the whole test
● A number of NM requirements are inherent to the system
● The speaker is audible to all present
For reference, the full test document can be found here.
Spring Validation Demonstration
The Spring Validation Demonstration will be held in Newell Simon Hall Lab B512. The necessary equipment includes the COBORG backpack, the vision subsystem, the voice subsystem, a computer to run the subsystems, a stopwatch, a ruler, the record sheets, a measuring tape, and a testing structure with shiny and matte finish. An area of roughly 10 feet in diameter will be required for the test.
First, the vision subsystem will be tested. The plate and camera will be fixed at specific locations, with a known relative pose. Then, a team member will place their hands on the plate and the camera will be signaled to determine its pose. The estimated pose will be compared to the ground truth, validating PM1.1, PD1.1, and PD1.3. The time from signaling to pose estimation result will be recorded, validating PM1.2. This test will be repeated for two other locations. The testing structure will then be moved to a different surface finish and then the test will be run again, validating PD1.2.
Second, the actuation and manipulation subsystem will be tested. The COBORG backpack will be fixed at a specific location and the testing structure will be fixed in a different location, including targets with known relative pose. Then, the backpack will be signaled to reach out to the target. The distance between the target and the center of the paddle will be measured, validating PM2. Then, the COBORG backpack will be signaled to return to compact mode. The maximum distance from the center of the first motor to any distal part of the arm will be measured, validating PM6. This process will be repeated. Then the process as a whole will be repeated for two different target locations to build confidence in the validation. Afterwards, a strength test will be performed, by signaling the COBORG backpack arm to hold up a representative panel, validating PM3.2.
Third, the voice subsystem will be tested. During the test, every trigger of the microphone will be recorded, as well as triggers that were supposed to occur, but did not. The microphone will be signaled to start listening. The user will say, “Coborg, move to goal” and record the subsystem’s response. The user will then read a randomly generated sentence. The user will then say, “Coborg, move to home” and record the subsystem’s response. The user will then read another randomly generated sentence. This process will be repeated twice to build confidence in the validation. Comparing the number of correct and incorrect responses verses triggers will validate PM4.1. Using appropriate commands will validate PM4.2 and PM4.3.
Finally, the nonfunctional aspects of the system will be tested. The COBORG backpack will be weighed on a scale, validating NM2. The COBORG backpack will be put on by all members present and their remarks about its comfort will be recorded, validating NM1.