This robot was created to be a scaled solar panel installation tool. The robot had to be able to place and remove panels at 3 different locations; a house with a roof angle of 45 degrees, a house with a roof angle of 25 degrees, and on a block parallel to the ground. In addition, the movements of the robot had to be controlled through the use of sensors and the implementation of a finite state machine.

Position 1: Staging Block

Position 2: 45° Roof House

Research and Design

I started by analyzing the gripper’s position in the three major positions; at the staging area, at the 45-degree house, and at the 25-degree house. From this analysis I was able to design a four bar to power the robots arm. The links of the four bar are highlighted below

Four Bar

Highlighted in yellow is the ground link (1.73 in), red is the driving link (4.36 in), green is the shortest link (1.11 in), and blue is the trailing link (5.92 in). Using the Grashof condition and the lengths of the links we can see that the design is a double rocker since the sum of the shortest and longest is greater than the sum of the remaining two links.

Gearbox

After the lengths of the Four Bar were determined, I next had to design a gearbox to power the linkage. Through extensive testing, I determined that when the motor was run at a PWM duty cycle of 100% the Max Stall Torque was 4.5 N. In addition, using the weight of the aluminum panel the max torque applied to the arm was calculated to be 0.551 N m or 55.1 N cm.

When calculating gear ratio of the gearbox I wanted to ensure that the motor stayed within 25% of the stall torque. This meant that the max torque the motor should experience is 1.125 N cm. Assuming that the gears have an efficiency 95%, since they were going to be 3D printed, and that I would have a 3-stage reduction the final gear ratio was calculated to be 1:40. I later reduced this to 1:45 in order to ensure that the stall torque was never reached.

I then preceded to create the gearbox and arm in SolidWorks. In the image below everything apart from hardware such as axles, bearings, spacers, nuts, and bolts were shown. Once the design was finalized all the parts were 3D printed and assembled.

Finite Element Analysis

To ensure that the 3D-printed gears would be strong enough to withstand the torque of the gearbox I ran a Finite Element Analysis on the smallest and weakest gear by applying 100N. The analysis determined that the gear's von Mises was 8.2 * 10 ^7 N / m^2 or 82 MPa. Considering that the tensile strength of ABS was 40 MPa, it can be calculated that the maximum force that can be applied to the gear is 48.8N. In addition, using the gear’s Pitch Diameter the maximum torque of the gear was calculated to be 30.9 N cm well above 25% of the blue motor’s stall torque.

Design Considerations

The initial iteration featured a driving arm that used a set screw and a live axle to control the four bar’s movement. However, while the arm was able to transport the lighter wooden panel, the arm began to slip under heavier loads such as the aluminum panel. In order to combat the issue the live axle was ditched and the driving link was redesigned to be connected the final driven gear.

Electronics and Sensors

To track and control the motion of the robot’s arm, encoders and a H-Bridge were used. Encoders allowed the motor to move the arm to a specific position by having the motor rotate a specific number of encoder ticks. In addition, the H-Bridge allowed the arm to move up or down because the motor could now be controlled in both directions, forward and back. Finally, the gripper was controlled by a servo and potentiometer.

The robot also features an IR receiver, an ultrasonic sensor, and a reflectance sensor. The IR receiver, located on the breadboard, allowed the robot to accept commands from the operator such as start, and emergency stop. The ultrasonic sensor detected the robot’s distance from the two houses or the staging area. The reflectance sensor, located under the ultrasonic, was used for line following when transitioning between the two houses.

IR Receiver

Software

The driving architecture was a Finite State Machine which controlled the robot’s movements and observed sensor data. FSMs transition the robot between various states depending on events that occur when the robot interacts with the environment. To view the final code see my linked GitHub.

The robot’s actions can be divided into 5 different tasks that needed to be completed sequentially: removing the aluminum panel on the 45° house (task 1), placing it from the 45° house, and placing it on the staging block (task 2), transitioning via line following to the 30° house (task 3), putting the wooden panel on the 30° house (task 4), and removing it from the 30° house and placing it on the second staging block (task 5).

Here are videos of the robot completing tasks 1, 2, and 3.

Software Considerations

While the robot completed all 5 tasks sequentially, the code could have been structured more efficiently. Due to how similar tasks 1 and 2 were to tasks 5 and 4 the code could have structured using the finite state diagram pictured on the right.

Using the state diagram for task 1

The robot starts in the Idle state. Once an IR signal is detected, the gripper’s servo starts. The robot then transitions to the Open Gripper state. The event in this transition is the detection of the IR signal and the action taken is starting of the servo. Once the linear potentiometer reads “open”, the blue motor starts and the robot transitions to the Raise Arm state. The next event checked for is the correct number of encoder ticks. Once reached the robot begins moving forward and the line timer is started. The robot is now line-following towards the house. As the robot approaches the house hysteresis is used. Hysteresis allows the robot to detect when it is too close to the house, the next event. The robot then starts closing the gripper and transitions to the close gripper state. Once closed the gripper then enters the turning state. After completing the turn the robot transitions back to the lining state as it transports the panel to the staging area. Upon reaching the staging area. The arm is lowered, the panel is placed by opening the gripper, and the robot returns to the idle state waiting to begin task 2. Drawing out the robot’s actions and labeling the events can be carried out for all 4 remaining events.

Final Thoughts

This was the final project of my 2nd Robotics Engineering course at Worcester Polytechnic Institute. The project was a deep dive into topics such as; mechanical design, control engineering, electrical engineering, and software engineering. Through the project, I was able to hone my skills in various software such as; MATLAB, SolidWorks, Visual Studio Code, Platform IO, Git, and GitHub.