This week, we worked mostly together as a group to complete the design report and test the bottle intake mechanism. Since the Amazon and AndyMark orders we had placed over the last two weeks had arrived, we were able to make a prototype of the spinning mechanism and test on a few plastic bottles. With the rotating rubber wheels placed onto the center rod with spacers in between, we attached the rod to a drill bit and simulated the speed at which the motor would have to spin the rod in order to actually propel the bottles up the ramp and into the storage area. We found that we would need to start with a high velocity before encountering the bottle in order to have enough speed. Otherwise, at a lower speed, the bottle would just be spun the opposite direction and move away from the intake. This is an aspect of the intake mechanism that we’ll have to fine tune.
Separately from testing, we directed the rest of our efforts towards our design report and through this process, streamlined our use case requirements again in order to incorporate a real-life requirement to base our own requirements on, resulting in a more practical 1 meter radius search area, and a 40 second search time for our metrics.
I believe we are on track for intake mechanism testing but must accelerate the development of our bottle detection algorithm. We need to fine tune the image processing in order to guarantee bottle detection accuracy at a 2 meter distance, and make sure we begin our navigation computations soon. I also figured out a more reliable way to compute the angle that the robot needs to turn to face the bottle, which is by using our cv object tracking to log when the center point of the object matches the center of the robot field of vision. I will be attempting to implement this next week.
0 Comments