April 29th, 2022:

Mechanical/Software Teams: After receiving the new power supply, the mechanical team wired this power supply to the electrical system inside the control box. The robot was then fully assembled and ready for use. Both the software and mechanical team then started constructing demonstrations for the final result of our project. This resulted in four unique processes for MATIE to execute including a 72 waypoint demonstration allowing the robot to stack six blocks on top of each other. This particular process took about 16 hours to construct with both teams involved. Once we had these demonstrations recorded, we began constructing the final presentation and poster for our Capstone. This involved retrieving all documentation we have taken in the course of this project, as well as new documentation such as a walkthrough of the GUI system. All this data was then compiled to these two documents allowing an audience to understand our design process and finished product. We then began constructing the research document and plan to finish it within the next few days. With this being completed, we will be finished with a successful Capstone project!

April 22nd, 2022:

Mechanical Team: An unnamed student, we'll call him ROMBOMBERTO, done broke dat power supplying thing. This was a big OOF on his part. This student is no longer allowed within 2.47 miles of MATIE (mostly due to his smell), so we feel this problem has been rectified. Following this supercilious remedy, the team finished the test document, as we had already completed the required system tests. Additionally, we received the controls box from RPL, painted it to better match the aesthetic of MATIE, and implemented it to our assembly, as can be seen below. This consisted of rewiring the motors and encoders through the control box's access ports, as seen in the Figure below. Once this was completed, the team began constructing the detail design document, as well as the detail design poster.

Software Team: Since the robot was out of commision this week, software team focused on writing a draft of the final paper to be presented to the Florida Conference for Recent Advancements in Robotics. They also worked on completing the final detail design poster as well as the Presentation for the capstone symposium

April 15th, 2022:

Mechanical Team: Though we planned on retrieving the controls box part this week, RPL has had printer malfunctions, causing our print to be further delayed to next week. So, instead of wiring our PCBs around the controls box, we continued running tests on the system to further build our test document. This mainly consisted of tuning each link by the software team, while the mechanical team was on standby to fix any problems that arise. Additionally, as the system tests were being performed, we began constructing the test document. Though this document was not finished by the end of this week, we plan to finish it early next week as all our tests should be completed then. Along with constructing the test document, the mechanical team constructed a table for the MATIE robot to sit on. This was done so the robot can be clamped to the middle of the table, instead of on the edge with C-clamps. This table can be seen in the picture below. Next week, we plan to finish the test document, implement the controls box, and hopefully start on the detail design document.

Software Team: The math for the inverse kinematics was derived using the new measurements and now allows the software to check student entered task space coordinates. The GUI was finalized and loaded with many different pre defined controllers that are meant for the students to use, as well as full implementation of student written controllers was added onto the GUI. The pick and place operation was also uploaded onto the GUI. Now a student can go onto the current GUI, pick either a predefined trajectory, and write a controller to test. Or they could enter in their own coordinates and pick a predefined controller to run the robot through those points.

April 8th, 2022:

Mechanical Team: To start off this week, we wired the PCBs to the motors, encoders, and power supply, though this will eventually have to be redone to run the wires through the controls box once its printed. This wiring can be seen below. With this done, and the software team's test code ready to be uploaded, we began running our first system test: testing the encoders' functionality. Through doing this, we discovered a problem with the encoder on the fifth rotation, as it was outputting incorrect values. After a long process of ruling out the cause of this issue, it was found that the inner bearing of the encoder was falling down to a space between the encoder and the link. We then designed a spacer such that this space can be exactly filled to make the encoder flush with the link, while keeping the encoder bearing held in place. Having implemented this spacer, the encoder values were then all correct. Next week, we plan to continue debugging the system so that system tests can be completed. Additionally, by the end of the week, the controls box was still not finished printing, so we plan to retrieve it from RPL next week and implement it to our assembly.

Software Team: The software team took about three meetings to test every link to make sure inputs and outputs were reading in correctly. Once every links decoders were reading in correct angles, and voltage outputs were sending the robot the correct direction, a full pick and place operation was written and tuned for the testing document. Matie had to prove it could be sent coordinates like the ones written, and actually operate the command.

April 1st, 2022:

Mechanical Team: The wire management clips were successfully printed, so we added them to the robot assembly. The Zero Jig was also finished, though in the wrong color. We then spray painted the Zero Jig black, waited for the paint to dry, then added it to the assembly. With the electrical team having finished constructing the PCBs, we verified that our designed control box fits all of the necessary components. We then took this part to RPL to develop a print plan, but we were informed that the box was just an inch too big in one direction for RPL's largest printer. We then devised a plan to redesign the controls box including a two story stair-step design, a wooden box design, and a box with some components mounted on the exterior. We decided that the third option was the best idea, having the pneumatic solenoids mounted on the exterior. The design was then finished, as seen below, and sent to RPL to be printed. Next week, we plan to temporarily wire the PCBs so that a full systems test can be achieved. Additionally, though highly unlikely due to a back up on the printers, we plan to implement the controls box if its finished printing.

Software Team: This week Sam soldered the decoder board which serves as a general purpose robotic interface for a raspberry pi. the Decoder board contains 9 encoder ports, 9 pulse width modulation outputs, and six outputs which can be used to control motor directions. Unfortunately, the GPIO budget of a raspberry pi did not allow for more outputs or inputs. Testing was done to ensure signals were correct. The Control architecture has written into the GUI to see if student inputs could be sent to the control script to move the robot.

March 25th, 2022:

Mechanical Team: This week, the mechanical team designed the zero jig such that it can bolt into the table mount part, as seen below. This part was then sent to RPL to be printed. We then sought out to implement some form of wire management for both our motor and encoder wires, as well as the pneumatic tube for the end-effector. To do this, we designed four unique clips to attach to preexisting holes on our robot such that they fasten the wires and tube to the links. These parts can be seen below. We then recalled that the specifications require the robot to lift a block with a width of 1.5", and our gripper currently has a max width of 1", so we created L-shaped 3d-printed components such that they could attach to the preexisting gripper and fulfill this specification.

Software Team: The PCBs came in and Sam spent the week soldering the motor driver board and conducting tests on each motor outputs and the end-effector circuit. Hunter wrote the first control architecture script that will allow the team to see if the robot could be zeroed based off Independent Joint Control. Huge step for the software team and as soon as the robot is wired, full system testing can begin

March 11th, 2022:

Mechanical Team: This week, the mechanical team finished construction of the MATIE robot, as all of the redesigned, 3d-printed links were finished printing. Additionally, we realized a design feature would be needed such that the encoders can be zeroed. Having discussed our options on achieving this, we decided to make a jig to attach to the back of the robot such that the end-effector sits in a unique position and orientation. Then, when the robot is sitting in the jig, we can manually set the encoders to the angle they are currently at, then manually move the robot to the zero position. We plan on designing this part next week.

Software Team: Sam worked on finalizing the PCB design to be sent out over spring break. This required a few redesigns to bring down cost and ensure that the circuit was correct. Hunter switched from computing the Robot's kinematic and dynamics terms in Python, to calculating them in compiled C code. This proved to be very fast and allowed for a 30Hz control loop sampling time. With this, a script to actually control the robot could start being written.

March 4th, 2022:

Mechanical Team: This week, the mechanical team designed a controls box to house MATIE's electrical components. To do this, we dimensioned each component that is to be stored in the box and created a CAD representation of these components. We then situated the components such that each output that needs to be accessible is positioned near a wall, while tying to arrange such that the controls box could be as small as possible. We then designed slots above each output so name plates can be inserted. This was done to make troubleshooting and design interpretation easy for the user. While only the mechanical team designed the controls box, both the mechanical and software team worked towards crimping wires this week. This included crimping wires for each motor and encoder, as well as applying pin adapters to each crimped wire so that it can be wired to it's associated component. Next week, we plan on assembling the MATIE robot with the redesigned links, given that they are finished printing.

Software Team: Hunter wrote a script to derive MATIE'S kinematics recursively. With this a function was made called Matie_Controller_Plugins that will calculate all needed components of a student written controller. Major speed issues are coming to fruition regarding the computation time needed to run through the robots main control loop. Running the python scripts as executables or possibly working c into the control might help mitigate these speed problems . Sam worked on the Printed Circuit Board and bench top testing the end effector circuit.

February 25th, 2022:

Mechanical Team: This week, the mechanical team redesigned the Base Link and Link 1. When initially thinking of a solution to this problem, we found that the motor shafts were bending due to a lack of support on one end of the shaft. To fix this, we designed a sliding plate to mount on the link, opposite to the motor. This plate would house a bearing such that a shaft can extend from the pulley into the bearing. This required redesigning the motor pulley's so that a shaft can extrude from it. Additionally, as recommended by the faculty advisor, we redesigned Link 1 to be one solid part, adding reinforcement bars across the gap. This redesign can be seen in the pictures below. With these redesigned finished, we sent in a print request for the Base Link, Link 1, both bearing mount plates, and both motor pulley's with extended shafts. Additionally, we sent in a purchase request for two additional bearing for the reinforcement system made for the Base Link and Link 1.

Software Team: The electrical team added the end effector into the pcb and started creating a circuit to make it open and close. The end effector is a pneumatic gripper that simply opens and closes at set distances. The software team ran into difficulties with processing time of calculating the system mass matrix, and vector of Coriolis and centipedal terms. Deriving the symbolic equations and running them as functions is proving too expensive for the Rasberry Pi 3. The robot will be ran off the Pi 4 but still this is problematic. The team is switching entirely to computing the dynamics of the robot recursively, which will hopefully cut down on computation time.

February 18th, 2022:

Mechanical Team: This week, the mechanical team finished building the dry assembly that was almost completed last week. While demonstrating this completed assembly to the software team, gravity did it's thing and caused the robot to fall over, causing the end of Link 4 to crack. To fix this, we decided to reprint the link, but with a redesigned end to help reinforce the face that cracked previously. In the mean time, the assembly will operate as intended, aside from a limited force we can apply to Link 4. After addressing this issue, we tensioned all the pulley timing belts on the assembly. While doing this, we noticed that, when both the Base Link and Link 1's pulley systems were fully tensioned, the axle of the motor started to bend, causing the belt to ride up against the pulley's wall. To address this issue, we sought out recommendations from our faculty advisor, though we eventually ran out of class time. Next week, we plan on redesigning these two links so that we can get a print request submitted as soon as possible.

Software Team: All of the motors have been tested and the data has been run. With this, the derivation of MATIE's dynamics can be constructed for full non linear robot control.

February 11th, 2022:

Mechanical Team: This week, the mechanical team fabricated the aluminum motor mount on the Table Mount link, as well as the aluminum base shaft on the Base link. To fabricate the aluminum motor mount, we milled the stock aluminum to reflect the overall shape of the part. Then, to make milling out the slots and fastening holes more feasible, we redesigned the part to have no radial cuts, then finished milling it. To fabricate the base shaft, we redesigned the part such that long bolts do not have to travel from the bottom of the base to the bottom of the Table Mount. This was done by merging the two separate shafts together and fastening it using a shaft collar on the bottom. To cut this part, several passes on the lathe was required, while a simple drill press with the mill was used to create the bolt holes on the top face. Once these two parts were fabricated, we were ready to start assembling the entire robot, as we received the encoders, motors, and belts from the purchase order. To do this, we made several trips to the Machine Shop to acquire hardware for fastening these links together. Once we acquired all the hardware, we began to compile each subassembly onto the Table Mount to produce the MATIE assembly, but were cut short by time constraints. We plan to finish the assembly next week.

Software Team: The software team ran system identification on each type of motor used in MATIE. The data will be collected and analyzed to find the motors viscous damping coefficient, coulombic friction coefficient, and the rotational inertia of the armature. These are vital to calculate the dynamics of our motors.

February 4th, 2022:

Mechanical Team: This week, the mechanical team made great progress on the construction of the robot. Each subassembly can be seen in the figures below. With the most recent Gantt chart update, the mechanical team plans to have the robot Fully constructed by February 18th.

Software Team: The software team built a circuit to read in encoder values and wrote code to turn the counts into joint angles. They also wrote a code to send out pwm, and directional outputs to the moto driver to control speed and direction of the motor. This is to run an experiment sending the motors random voltages and recording the angular response of the output shaft.

January 28th, 2022: Hunter and Sam finalized the layout of MATIE'S Graphical User Interface. They also decided .cvs files would be the default format for imputing orientations and positions of the robot. With this, a program was written to index through the file input to grab all the x, y, z positions as well as orientation and claw status( open or close). The next task for software is running the motor experiments for system identification. Every 3d part and most ordered parts came in this week. Robbie found an error in one of the 3d printed parts and re submitted it for printing. After that, Robbie and Landen went straight to getting the robot ready for a dry assembly. They spent most the week milling the aluminum piece that would be fixed onto the table mount of the robot. The next task for the mechanical duo is finishing all the initial modifications needed to assemble the robot correctly.

January 21st, 2022: Hunter and Sam started creating a GUI for the user to enter and select data. This program will run on the Raspberry Pi and allow for direct control of the manipulator. The mechanical team placed the full parts order and have started to receive select components.

January 14th, 2022: This was the first week of our spring semester. Our team didn't do much over winter break but Robbie was able to add an end effector to MATIE and finalize the CAD files. This week was used to create a more in depth project plan outlining the work that was needed and their respective due dates.

December 12th, 2021: As a team, MATIE members came together to write the final Preliminary Design Report, and Presentation to end off the semester. MATIE presented the project on December 12th to a panel of previous Robotic students and Professors

December 3rd, 2021: Sam Esse started working on the Printed Circuit Board (PCB) needed to implement all electrical components for the system. Hunter Smatla made an open loop simulation of MATIE, and started the closed loop controller needed to send MATIE to a pick and place operation. Robbie and Landed finalized revision one and started on revision two for the 3d model of MATIE. Adding all necessary supports and features to the Robot to allow for a successful print and usability.

November 19th, 2021: Landen and Robbie finished revision 1 of the CAD model for MATIE. A photo of the CAD can be found under the robot tab. The mechanical team also calculated the mass of each link by consulting the Rapid Prototype Lab on 3D printing. Since the CAD was finished we had the values for each link which allowed us to create an accurate forward kinematic function. We also got started on the simulation of the robot using MATLAB. Hunter and Sam parametrized the motors which allowed for the simulation to include motor dynamics.

November 12th, 2021: The CAD model is almost done. It should be complete by next week so that we can all come back together and work on creating the Equations of Motion (EOM). Hunter Smatla and Sam Esse created program functions on a Raspberry Pi that calculate the forward kinematics for M.A.T.I.E.. The function can not be completed until the final measurements have been finalized by the Mechanical Team . Awaiting the CAD, the Software team switched gears into parameterizing the motors that Robbie Shaw and Landen Rhodes had picked out for the mechanical design. These parameters are essential for the (EOM) used for creating a dynamic simulation of M.A.T.I.E.. The mechanical team has made significant advances in the CAD model. Landen Rhodes designed a table mount for M.A.T.I.E. That's secures the base to the ground and allows for the base rotate. He also designed a system on the mount to allow for tensioning of the belt on the motor. Robbie Shaw 3D modeled the base of the robot and links one and two. He also designed a tensioning system in the link that allows for the belt to be tensioned. This mechanical feat will be used in the next two links that rely on motor driven pulleys. Next week the Software team plans to finishing system identification of our motors, and will move on to creating a recursive dynamics function to aid in EOM. The mechanical team plans to finish the 3D model of the robot and start creating the .stl files for a proper dynamic robot simulation.

November 5th, 2021: The team is now working in a split format. We've officially been turned loose and aiming to maximize work getting done, we separated into a Software and Mechanical team. Robbie Shaw and Landen Rhodes are spearheading the mechanical design while Sam Esse and Hunter Smatla work on software aspects of the robot. This week, the mechanical team finalized a mechanical design, and created a complete parts lists needed to build it. They also completed a bill of materials which serves as an estimated cost for the robot. The software team researched MicroControllers and started working on creating functions to solve the forward kinematics of the robot. The entire team also worked together to complete a project plan that will drive the project for the rest of the semester. Next week the mechanical team aims to complete a full CAD of the robot, while the software team aims to complete the robot's forward kinematics, and also the derive the kinetic energy equations of the robot.

October 29th, 2021: This week we took our peer reviews of our specifications document and created another revision. The most recent rendition is located in our Documents tab. After we were satisfied with our specifications document we sent on our own to create our final conceptual design. First we created a schedule so that we can hold ourselves accountable and then we assigned tasks to each person biased on their strengths and what they wanted to do.

October 22nd, 2021: Specifications Document Presentation: The team presented their robot's specifications to the class and once again received critical feedback how to improve. This was a very dark time for M.A.T.I.E. as most of our specifications brought disgust and anguish to Dr. Isenberg. Pulling together all our brainpower, and using all the classes suggestions a finalized specifications document was created. You can find this paper under documents as "Project M.A.T.I.E's Specification Document"

October 15th, 2021: Presented Group Conceptual Design. Again we went back and revised our document with the feedback that we received. Once we had finished the revisions we were given instructions on specifications and how they fit into the design process. Requirements tell us what the robot needs to do. Specifications tell us how our system will accomplish the tasks.

October 8th, 2021: Working on team conceptual design. Pulling all the ideas from individual designs, a morphological chart for project M.A.T.I.E. was created. This is a chart that displays certain needs of the robot like power transmission and end effector, then lists all the possible ways we could accomplish that certain task. After this, all the ideas were taken from the morphological chart to build decision matrices, weighing all the ideas across certain criteria to find the objectively best one for our purposes. The team used this decision process to build the final conceptual design and then created a document explaining our design in detail. This paper can be found under the documents section of our website as "Project M.A.T.I.E.'s Group Conceptual Design". The team then made a PowerPoint to present the design to the class and practiced for the October 11 presentation date.

October 1st, 2021: Present Individual Conceptual Designs. This week we presented out individual conceptual designs. Shown below are each of our conceptual CAD drawings. We all received feedback on our presentations and it allowed us to see which aspects we should keep for our Group Conceptual Design and which areas need improvement before being added to the Group Design.

September 24th, 2021: Literature and Patent Search Presentation. The team compiled their individual findings into one document and presented it the class. After using the class feedback to revise the document, the initial research portion of the project was done. With initial research out of the way, the team could start Individual Conceptual Designs. Each member of the team had to come up with a different idea of how to make an educational robot. This includes designing the mechanical system, an overview of how the software will operate and a basic description of the electrical system.

September 17th, 2021: Requirements Document Presentation. The team presented a draft Requirements document and received critical feedback from the class. Using this feedback, the team revised the Requirements document until we were satisfied. The final Requirements document can be found under the Documents section of our website. We also received another short briefing, this time about the Literature and Patent review. This is where each member of the team individually searches for patents or projects relating to educational robots. Most problems in engineering have already been solved, so researching other people's solutions can be a valuable resource in helping us design our bot.

September 10th, 2021: Requirement workshop. We received a short briefing on the proper way to make requirements and then we set out to create a requirements documents for our robot, M.A.T.I.E.. Requirements are needed to specify the functions the system needs to fully solve the problem presented. This is a vital step in the design process as good requirements can drive the course of a project and deters the team from deviating from the main goal of project.

September 3rd, 2021: Project Team Created. The Team consists of Landen Rhodes, Robbie Shaw, and Hunter Smatla, studying Mechanical Engineering, and Sam Esse studying Electrical Engineering. These lucky few joined forces to create a dexterous manipulator to aid in future robotics labs.