Chassis

Wheel System:

The robot uses a differential drive system and the chassis is designed to have the wheel axle going through the geometric center of the robot, so the robot has a minimal turn in place footprint. 

Because the drive wheels are on the geometric center axis, the robot needs supports on both front and back of the wheels. Due to the intake mechanism at the front center of the robot, the front two corners of the robot are supported by casters. Because the rear caster of the robot should have as little gap to the ground as possible to avoid the robot rocking, but it also can't be too low to the ground, making drive wheels lose contact with the ground, the rear caster is designed to be spring loaded. It uses a piece of duron as a leaf spring and has a neutral position below the drive wheel contact patch to achieve a balance of chassis stability and good drive traction. 

To save space and minimize mechanical complexity, the drive wheels are directly connected to the gearbox output shaft of the motor, and the motor is mounted directly to the chassis plate. We don't expect a lot of load on each wheel since we have three casters to spread the load, so the gearbox should be able to bear the axial load the wheels transfer to it. 

Structure:

For the chassis to be stiff and the drive system to perform well, the motors should be very rigidly mounted. Therefore we designed strut towers to transfer the bending load that would be imparted on the chassis plate, into a compressive load of a plate above the chassis plate, similar to a strut bar in a performance car. 

The chassis also needs to support the weight of all the components mounted on the plate above it, so there are several supports all over the area of the chassis plate where loads are expected. The front part of the plate above is supported by the elevator ramp and the front casters, the center is supported by the strut towers and side panels, and the rear is supported by the rear panel. 

Navigation Sensor Mounting:

The robot needs to have room for ideal sensor mounting positions, so their location have been planned around. 

The ultrasonic sensors are expected to be at the side that has the intake mechanism, so that's the front. However, it's very important that the ultrasonic sensor be at the right height. They need to be high enough so they don't pick up reflections produced by stray ACORNs on the field, but they also need to be low enough so that they can still detect the field walls, which is critical for alignment. To solve this, we positioned the ultrasonic sensors at the right height, and designed wave guides for them, so the ultrasonic waves never propagate to below the sensor's height and bounce back due to stray ACORNs. 

Besides the ultrasonic sensors, our robot also featured 3 reflectance sensors to detect the tapes on the field for guidance purposes. Two sensors are at the front, spaced 1.5 times the width of the tape apart, for line following purposes. The forward positioned sensors ensures tracking stability. A side sensor is right on the wheel axle, so the robot would know when it has crossed the center point, thus able to turn in place and continue to follow the tape.

ACORN Collection

The Ramp:

The initial plan was for the robot to contain two large magazines 44cm long which would begin at an angle and swing down into a level position after the start of the round. They would be able to hold 12 acorns between them and would be loaded from the trees using an arm. The firing mechanism was a rod with multiple rubber bands attached that was held in place by a quick release mechanism. There were two issues with this design. First, it only allowed the robot to fire once, which would ideally be at the end of the round. The second is that it took up over half of the internal volume of the robot and didn't leave enough room for motor and wheel mounts. 

The critical innovation that allowed for the pivot away from this design was the elevator ramp.

In theory, this ramp allows for multiple acorns to be stacked on top of each other by simply pushing each successive acorn up the ramp. The stacking motion was achievable because the acorns remained level with the ground as they ascended the ramp. This was because of an angled cylindrical cutout in the ramp which made even contact with all points along the circumference of an acorn's bottom surface simultaneously.

The Magazine:

As for the magazine into which the acorns were loaded, the first geometry tested was a half-pipe. This magazine geometry demonstrated the importance of walls in keeping the acorns aligned, but there were binding issues which would cause acorns to become stuck if they ever became angled toward the back of the magazine. This led to the second and final V shaped magazine geometry which removed the material the acorns were binding against. This design also moved points of contact between the magazine and acorns closer to the acorn's geometric centers which reduced rotational moments and further reduced the likliehood of binding.

The Intake Arm:

The issue with the elevator ramp is that it requires the intake arm to be tall enough to make contact with acorns both on the ground and as they are moving up the ramp. There is another conflicting design constraint, however, which is that the arm needed to be short enough fit within the 4cm gap between the the tree and the ground

There are two obvious solutions to this. The first, which would have been much more effective, is to have two arms: one short arm to reach under the tree, and one tall arm to push acorns up the ramp. The second solution, which was implemented on the robot is to have one arm with a variable height which can fit under the tree at one point in its movement arc and reach the top of the ramp at another. The plan in that case was to have a system of two pieces of anchored fishing line traveling through various channels along the arm in such a way that one piece is tightened while the other is loosened as the arm rotates from the open position to the closed position and vice versa. These pieces of fishing line were attached to two sides of a hinged finger element on the arm which would swing from flat to vertical almost immediately at a very specific stage in the arms movement. This system was more or less functional at showcase, but the amount to fine tuning it took to make it function was too high.

ACORN Launch

The ACORN launch system functions by striking the ACORNs from behind using a steel firing pin. The firing pin is propelled by a loaded rubber band, and the firing pin is pulled back after firing by a snail drop cam powered by a worm gear motor, thus achieving the repeating firing action. During typical operations, the system can sustain around 1.6 rounds per second. 

The firing pin assembly is cylindrically constrained by a linear shaft bearing, and the rotation is constrained by the wings on the rear of the firing pin, thus the firing pin only produces linear motion

The firing system also features a limit switch to let the robot know when the mechanism is fully wound back right before firing. This position feedback is necessary for the firing mechanism to clear the way for more incoming ACORNs after firing to avoid jams.