CMU Mechatronics 2013 - Team E

Progress Reports

Update 05-06 - Public Demo

This week the team had its public demo! We presented our awesome video, seen above, and we took second place in the disc-bot competition. This was a satisfying conclusion to a good semester of work. Check out Dr. Dolan's comments on our robot!

Update 04-24 - System Demo 7

This week was the last system demo before the final demo.  We had many improvements made to the robot to show off, most notably being a re-design of the entire mechanical structure.  Central to the design, the hopper has been re-designed to take two pairs of belts, as opposed to three individual belts, and it is rectangular as opposed to triangular.  The hopper was fabricated and assembled, but was not incorporated electronically into the robot in time for the system demo.  In addition, the track was re-cut reducing the bend from 90 degrees to 45.  This was done to economize on space and weight, because we did not need the entire length of the track to accelerate the frisbee, allowing us to make the entire top platform narrower.  Strategic reinforcement of this area ensured that we would have fewer issues with the structure of the robot flexing and sagging as it had earlier.

CAD model of the revised, shorter track, and the new hopper (walls facing viewer are transparent for visual aid)

Update 04-17 - System Demo 6

This System Demo saw an attempted implementation of the motorized belt hopper, which due to its elaborate and somewhat fragile construction was unable to withstand being stepped on by a clumsy team member, and was unfortunately presented in pieces at the demo.  This demo served less as a demonstration of progress and more as a demonstration of what still needed to be done on the mechanical end - construction of a structurally sound hopper, integration of that hopper into the shooter platform, and reinforcing of the robot, as it is prone to flexation, vibration, and large amounts of mechanical slop.

Update 04-10 - System Demo 5

The fifth System Demo saw the robot take another substantial step forward.  The yaw and pitch subsystems were modified, and took another step towards completion, after the motors for each system were specified.  Instead of using the two DC motors, we chose to use the servo for yaw and the stepper motor for pitch.  These choices were made because it allowed us to have an easier job of determining and holding our x/y aim position.

Update 04-03 - System Demo 4

This week Neil worked on an input and output panel for the robot. In addition to the e-stop that we had already, we now have an init button that calibrates our machine, either when the robot is in the idle state and needs to be calibrated for the first time, or from the ready state and needs to be recalibrated. We also have a start button that starts the firing sequence from the ready state. We have switches to control which test the robot is running, and we have status LEDs that report on the power state of the psu, whether the estop has been hit, and the current state of the robot .

Input Panel

Jerry and Qi finished the designs and laser cut the pitch and yaw subsystem. It looks great and the lazy susan moves smoothly. Ben assisted with the assembly of these parts.

This coming week we will wire the sensor and motor lines into the robot and have some real pitch and yaw movement happening. We will be rewriting our code slightly to use a servomotor for yaw and our stepper motor for pitch. Hopefully our wheel motor will arrive and we will have the parts ready to attach the motor. Then we can try to shoot discs with our launcher. We will also be adding an MDF bottom to the robot for stability and electronics housing.

Update 03-27 - System Demo 3

This week we worked on integrating actuators and sensors with our final state machine code, finalized mechanical designs, and began fabricating parts for the launcher and hopper. The ECEs took our object classes, primarily written by Ben, and began calling their functions within our state machine to actuate motors and receive data from sensors. This involved physically wiring the devices to the appropriate pins, observing the behavior of the system, and debugging if performance was not expected. We found and corrected several bugs, especially regarding attaching interrupts. 

We have debugged and have examples of the following classes: states, motor, encoder, limitswitch, power, simplestepper. The CMUcam4, CMUcom4, and target classes were tested last week. We have examples of the physical devices corresponding to these classes strapped to a piece of foam board until we can test with the devices attached to mechanical systems. Our entire system no longer needs a power supply or a computer to run, it all works with the ATX power supply.

Jerry and Qi finished up cad models of the hopper and the launcher and attempted to fabricate these pieces, but had trouble with laser-cutting the MDF board. They were able to produce the launcher track and the hopper.

System Demo 3

Update 03-20 - System Demo 2

Fresh back from break, Neil and Ben worked to get the CMUcam4 and the sound effects working and also assigned pins for the Arduino Mega. We demonstrated this during the system demo by showing how the system could track different colored targets and respond to changes in the target arrangement  demonstrated by different sound effects. Jerry and Qi showed their updated launcher design.

Qi and Jerry will finish fabricating the launcher by Friday afternoon and have a CAD design for the hopper by Monday. Neil and Ben will get more sensors and actuators connected electrically and further refine and debug the software.

System Demo 2 Setup

YouTube Video

Electrical Connections

Updated Solidworks

Update 03-04 - Midsemester Presentation

Update 02-29 - System Demo 1

This week we designed the overall state machine for our robot, further developed our software libraries, and continued working on our Solidworks designs. Neil designed and programmed the state machine for our thrower. We broke the machines operation into several states: E-stop, Idle, Initialize, Ready, Scan, Schedule, and Fire. Ben implemented tons of code for the sensors, motors, and objects. Neil helped him out with the overall pseudocode and made one of the libraries (Schedule). We made the design decision to use an Arduino Mega instead of our original proposal to use an Aruino Due, and placed our order. Qi and Jerry have been making Solidworks designs for the final system, which are beginning to really show the shape of the robot. We have all been working together to weigh the pros and cons of different implementations of our subsystems.

Overall Design

Yaw Control

Update 02-22 - State Machine Lab

Team E State Machine

This lab report shows the progress of our team project since February 13th. This week we designed a state machine to show the capabilities of a subsystem of our machine, further developed our software, and continued working on our mechanical designs.

Our yaw and pitch “calibrated” by hitting a limit switch on one side to establish a reference point. Then a Frisbee could be eject from the hopper (Fig 3) and moved by the wheel. 

Neil designed and programmed the state machine for the demo and modified the prototype to use a new sensor for its calibration input. Ben worked on driver files for interfacing with motors and sensors and helped with the state machine. Jerry mounted the motor and servo to actuate our subsystem for the demo. Jerry and Qi worked together on their Solidworks models and parts orders.

This week Ben and Neil will try to lay out as much driver software and high level software as possible, and write towards the middle as we receive and test our parts. Qi and Jerry will be finishing the first Solidworks drafts soon.  

Update 02-13 - Sensors Lab

This week we integrated sensors into a motor rig, translated analog output of sensors into usable data, and demonstrated how to use this data to affect the performance of our machine. We also began making solidworks models of our machine.

Sensors included a potentiometer, a rangefinder, a photogate, and an ambient light sensor. Every sensor but the photogate was fairly easy to implement, requiring just ground, 5V, and an output line. The photogate required particular resistors to properly set the LED and the phototransistor. 

Neil and Ben worked together to complete the sensors lab, with Neil focusing on wiring the sensors and Ben focusing on writing the code, but each helping the other. The ECEs also taught the MechEs about sensing technology and the structure of the code. Qi and Jerry have started designing the mechanisms in solid works (Fig. 5) and are working on their parts list.

Subpages (1): [Untitled]