Project SIXT33N


Music Version   |   Speech Version



It’s time to show the world what you’ve learned in the EE16 series! For the final project you (and your partner) will gather all that knowledge and skill to build a flavor of the SIXT33N robot (no, we’re not leets. Read: 6T3-3N).


The SIXT33N is a mobile robot on 3 wheels (2 drivable) that moves around according to some input. It uses the MSP430 Launchpad as its guts with some circuitry for driving the motor and sensing through a microphone. It also runs on a 9V battery, so you don’t have to chase it around as it moves. There are 2 different flavors of the project, each described below:


Version A: Music Recognition

In this version, the SIXT33N will recognize two different genres of music. SIXT33N will turn, drive forwards, and change speeds, depending on the genre and loudness of the music that it hears.


Version B: Speech Recognition

This version of SIXT33N will recognize 4 different voice commands. It will then move forward, left or right based on those commands.


The version that you choose is up to you! Both versions have a similar amount of work, and have three main sub parts: circuits, PCA, and controls. For the first two weeks, half of your team will work on the microphone front end circuit, and the other half will work on the PCA classification. In the third week, one person from each group will work on integrating the front end circuit with the classification, and the rest of the team will begin implementing the controls. Finally, you will bring it all together into one glorious machine.

The timeline for the project is roughly as follows:

Date Circuits Group PCA Group
03/28 Microphone Front End PCA Classification
04/04
04/11 Processing Integration Controls
04/18
04/25 Final Integration
05/02 Demo


Even though most of the phases span 2 weeks, we will have checkpoints every week so each group is progressing together. These checkpoints will be clearly labeled in the iPython Notebooks. The checkpoint for week 1 is due in the beginning of your lab in week 2. Each checkpoint is worth 5 points, and late checkpoints are worth 4 points. On top of these checkpoints, there is a short final writeup detailed in the last iPython Notebook.


Phase Checkpoints
Circuits Week 1: Circuit schematic agreed by GSI
Week 2: Working microphone front end circuit
Data Processing (PCA) Week 1: First pass through PCA with sample data; GSI feedback
Week 2: Classification target met in Python
Controls Week 1: Open loop parameters gathered, derived method to set desired eigenvalues, plan for turning
Week 2: Car drives straight and turns
Processing Integration Week 1: Take new data and run through PCA, show GSI results in Python
Week 2: Classification reasonably accurate in Launchpad / LEDs responding to input
Final Integration The final product!

This project is meant to give you design experience and creative control. Once you have implemented the basics version of SIXT33N, we encourage innovation. Any project that goes beyond the requirements will receive extra credit!


Phase Version A: Music Version B: Speech
1 Mic Front End and Signal Processing
  • Design microphone front end schematic
  • Build circuit
  • Read ADC output in PC
2 Data Processing
  • Use class microphone to record sample music genres
  • Examine spectrum of different genres
  • PCA + Classifier (2 genres)
  • Check accuracy
  • PCA projection on Launchpad
  • Use class microphone to record sample voice commands
  • Examine envelope of different commands
  • PCA + Classifier (3 commands)
  • Check accuracy
  • PCA projection on Launchpad
3 Processing Integration
  • Collect robust data set using front end circuit
  • Re-run PCA classification
  • Integrate final PCA into Launchpad
4 Controls
  • System modeling
  • Eigenvalue placement
  • 2D movement simulation
  • Construct car
  • Motor driver circuit
  • Wheel encoder circuit
  • Move at constant speed + direction
5 Integration
  • Control speed using loudness
  • Incorporate PCA projection - only move and turn when listening to a genre
  • Control speed and direction using PCA projection