The Project

The project for ME218B involved collecting Nerf balls from a central dispenser and then placing them into one of three different goal areas, for a varying number of points. The full project description can be found here.

The Strategy

Our strategy revolved around a robot which would drive to the beacon dispenser, orient itself, and then release a latch between two halves of the bot. The front half of the robot, containing most of the sensors and the drive train, would drive towards Goal Three, pulling with it a tape measure to act as a bridge between the two halves of the bot. The back half (“backbot”) would remain at the beacon dispenser, continually requesting balls and funneling them onto the bridge between the two halves. The balls would travel down the bridge and fall into Goal Three, where the geometry of the front half would aid in stacking the balls.

Perceived Advantages

The primary advantage of this robot design was the very high scoring potential. Due to the relatively small size of Goal Three, we felt that, with the aid of the front back extending the box, we could stack significantly more balls in that goal than a robot without a stacking aid. As Goal Three provides the largest number of points, being able to deliver more balls into Goal 3 than any other team should be a tremendous advantage.

Other advantages include:

- the sheer novelty of the design. We tackled issues no other team had to face.

- Less sensing required than a team which ferries balls from the dispenser to the goal

- Simplicity of control scheme. While this never really bore out due to time constraints, our robot’s behavior could be written almost purely sequentially.

Perceived Disadvantages and Challenges

The main disadvantage to this design concept (that we thought of at first) was the shear difficulty of it. We faced significant mechanical challenges in latching the robot halves, allowing them to unlatch cleanly, causing the backbot to stay in place effectively, allowing the bridge to swivel appropriately, pulling the tape with the front half, and doing all of the above within the space constraints given.

Performance

The performance of 50 lbs of Robot Love bore out the potential of our strategy as well as the effects of having a very difficult design.

In testing, we were eventually able to successfully score the maximum number of points (96) by funneling all 20 balls into Goal Three. We successfully found the ball dispenser, delatched the two halves of the bot, found Goal 3, placed the stacking aid, and begin the reception of the balls.

However, in competition, we suffered due to lack of robust code and a further testing period. Our tape sensor picked up a false positive while finding the right angle bend in the black tape strip, and began its beacon-approach-and-detach routine much too early. If we had had more time to calibrate our tape sensors, or to write code that could validate a tape sensor hit through other means, we believe that we could’ve had a lot of success.

Lessons Learned

Along the way, we also ran into several unforeseen issues. The most severe issue we encountered was most surely weight. To counter the rather large moments placed on the bot by pulling the FatMax bridge, we had to place upwards of 20 pounds of counterweights over the casters of frontbot and at the back of backbot. The increased weight limited our speed, and the distribution thereof also limited the force we placed on our wheels, limiting our available driving force. In addition, the added weight made latching and de-latching much harder, as we were more prone to binding. We initially used set screws in our motor mounts and in our shaft couplers and we learned two facts about using set screws. First, flat head set screws should be avoided—it is very difficult or impossible to put a lot of torque onto a flat head set screw without breaking the screw, which means the screw may frequently work itself loose. Second, set screws on especially hard shafts are not as effective—it is just as easy to use a pin, but the pin will hold while the set screw may slip.