2.3 Design Process

Preliminary Design Ideas: 

Our initial design consisted of using a quick return mechanism, pictured in Figure 1 below, due to it quickly building up velocity, creating a good shot path. However after comparing it to a four-bar mechanism, like Figure 2 below, we realized that that four-bar would be better. This is due to the four-bar having rotational bearings that are better than prismatic bearings because prismatic bearings would have to fight more friction than rotational bearings. This friction would hinder the ability of the mechanism to reach it's best escape velocity. The slotted mechanism on the quick return would also have less mechanical advantage than the four-bar, leading to a overall better escape velocity on the four-bar mechanism. 

Figure 1: Quick Return Mechanism

Our second preliminary design idea was to use a 4 bar mechanism to launch the ball with a driving link that would have the end effector link rotate about a point and launch the ball directly up and forward without "dipping" the ball backward and then launching it like a catapult would. In Figures 1 and 2 the three different S-Curves are the shot path of various basketball players, with the far right shot being modeled after Steph Curry. Due to both the simplicity of the shot and the accuracy of Steph Curry's three-point shots we wanted to pursue that model path. As this was a rudimentary sketch to determine the initial path of the ball as it was launched, more work had to be done to calculate and detail the specific force and velocity needed to produce an accurate shot. From this initial model of the shot path, we were able to determine the type of motion that would provide us with the most accurate shot. We determined that the shot path could not have any wasted motion. As seen in the figure below, other launch paths, such as that of a catapult, include an initial "dip" that brings the ball down first before it launches which prevents the ball from achieving an optimal path.

Figure 2: Initial Four-Bar Linkage Mechanism to Model Shot Path

Prototype and Iterations:  

Based on the constraints and requirements we were able to shape from our preliminary design ideas, we designed our initial prototype. Our initial prototype, shown in Figure 3 below, is a four-bar linkage with about 60 degrees between our two ground links. The end effector has a spoon attached to the fourth link to replicate the hand/cradle the basketball will be placed into for the final prototype. Using link 2 as the driven link we were able to create a good launch arch and escape velocity, proving to us that we'd be able to refine the linkages to have less play and be able to create a consistent shot into a basketball hoop with the use of a motor. 

Figure 3: Initial Prototype

Final Design: 

Our final design for our automated basketball shooter ended up using the 4-bar linkage from our prototype with a 100 RPM 12V Motor to turn the driving link which would eventually propel the ball forward from link 4. The launching mechanism, as shown in Figure 5 below, was placed on an elevated court that is two feet long, which is shown in the full picture of our basketball shooter in Figure 4. The court includes four spots, marked by velcro strips, to place the basketball hoop so that the user can shoot from different distances. Link 4 includes a ball holder that hugs the bottom half of the 25 mm basketball, ensuring that the ball is released at its peak velocity along the trajectory. An L298N Motor Driver and an Arduino were also included to control the motor that would drive the whole mechanism. A button system to shoot and cycle through different shot distances was also included for ease of use for the user. 

Figure 4: Full Final Design of Automated Basketball Shooter

Part of the design process for our final design was finding a way to assemble the launcher for the court. First, we needed to figure out how we wanted to attach the motor to link 2. Shown in Figure 6 is the motor mount we 3D printed. The motor mount mount allows the motor driver to be screwed directly into the mount using heat inserts on the side face. Initially, we wanted to screw the face of the motor directly onto the motor mount, however, we quickly realized that it wasn't feasible as it wouldn't allow the shaft of the motor to reach the opposite side of the ground link. We then went through a few different iterations of press-fits to securely mount the ground link of the launcher to the basketball court. Below in Figure 7 are the iterations of press-fit hole sizes that we used to determine the best mounting. Additionally, we had to iterate on the design for the ball holder to both hold the ball as snugly as possible so that it can launch properly and so that when it is attached to the end of link 4, the ball holder is not cantilevered and adding more stress to the joint than necessary. In Figure 9, the two iterations of the ball holder design are shown. The design we ended up with attaches the holder to the end of link 4 by being bolted in on either side of the link rather than to one side of the link. Another design aspect that we iterated on before settling on the final dimensions, was the fitting for the D-shaft of the motor. In Figure 10, the different hole sizes that were tested for press-fitting are shown. A proper press fit here was crucial because if the D-shaft had any wiggle room it would cause the links to not rotate as calculated, making it very difficult to get an accurate shot trajectory.  When testing we found that the 3D-printed basketball balls often bounced out of the 3D-printed hoop. To solve this issue we attached velcro strips to reduce some of the kinetic energy on impact helping the balls stay within the hoop.  The strips also helped ensure that the locations would be consistently placed with various users so we decided to keep them in the final design, rather than needing to know the lines on a basketball court.

Figure 5: Final Linkage Design with End Effector and Motor Setup

Figure 6: Motor mount with attachments for motor driver

Figure 7: Press-fit Size Iteration for Mounting

Figure 8: Ground link attached to the base plate using wood glue

Figure 9: Final ball holder (left) and initial ball holder (right); modified so the end effector isn’t cantilevered

Figure 10: Iteration of press-fit for D-shaft of motor