+ All Categories
Home > Documents > Wiimote/Kinect Lab Design Senior Design December 2011, Group 16 Adviser: Dr. Tom Daniels Brenton...

Wiimote/Kinect Lab Design Senior Design December 2011, Group 16 Adviser: Dr. Tom Daniels Brenton...

Date post: 24-Dec-2015
Category:
Upload: simon-harrison
View: 215 times
Download: 0 times
Share this document with a friend
Popular Tags:
31
Wiimote/Kinect Lab Design Senior Design December 2011, Group 16 Adviser: Dr. Tom Daniels Brenton Hankins Rick Hanton Harsh Goel Jeff Kramer
Transcript

Wiimote/Kinect Lab DesignSenior Design December 2011, Group 16

Adviser: Dr. Tom Daniels

Brenton HankinsRick HantonHarsh GoelJeff Kramer

Outline• Problem/Solution• Requirements• Risks, Costs, and Schedule• Wii

• Hardware• WiiWrap and Software

• Kinect• Hardware• Software

• Test Plan• Project Status• Work Distribution• Future Plans

Problem Statement

The Wii Wrap software used by freshmen Computer Engineering students is simple but needs improvement for installation and use by novice users such as a high school science teachers. In addition, similar software should be developed that uses the Microsoft Kinect sensor as input to student-written programs.

Solution

Solution

Solution

Functional Requirements• Develop an installer for the Wiimote project.• Identify and Remove errors from the Wiimote lab software • Correct buffering problems and synchronize user program with

the incoming data stream.• Make Kinect sensor data streams available to the user in an

efficient and real-time manner• Provide an interactive application that students can program

using the Kinect data.• Have an interface to write programs on the received Kinect

data and run them.

Non-Functional Requirements• Reliability: We will refine and package the software that is

designed. This pertains to optimization of real-time processes such as data processing. A testing plan shall be developed for this purpose.

• Portability: The project shall be portable between computers. We will cater mainly to windows, but will ensure that our libraries are extendable in case there is a need to adapt it to another system. Additionally, the software shall be easy to install and remove.

• Usability: The project shall be easy to use, both for instructors and students. It shall simplify tasks to set up and run projects by means of intuitive GUIs so that students can focus on problem solving aspects as much as possible.

Assumptions• The software is designed to be used in a Windows computer

lab that has both a bluetooth dongle and a Wiimote or a Kinect sensor at each computer terminal.

• We are currently assuming that the Microsoft SDK will be released in the coming months, otherwise development will continue based on the OpenKinect software on the Windows platform

What makes us unique?• Using video game hardware for to teach programming in an

educational setting• Designing software for students with little programming

experience to work with 3D video capture & processing functions

• Interactive API and GUI that makes it easy for students to program without worry about OO programming or managing system/hardware issues.

Existing Developed Software• The Wii Wrap software was previously developed by a student

at Iowa State• Similar software for the Wii Remote exists such as Wiimote

Whiteboard, WiinRemote, or GlovePie• OpenKinect is a package of software created by enthusiasts to

run Kinect programs on Linux, Mac, and Windows using a variety of languages

• Within a few weeks Microsoft will release a Kinect SDK specifically for Windows

RisksRisk 1: Temporarily lose a developer from our team for unforeseen reasons.Mitigation 1: The team will do our best to continue work without the developer while they

are gone and rely on a detailed documentation trail. We will use technologies like Skype and email to keep in touch with the developer during their absence.

Risk 2: Unable to keep tasks fully on schedule due to other projects and schoolwork.Mitigation 2: Set up minor deadlines to make progress on larger tasks. Risk 3: Important equipment becomes lost or nonfunctional (Wiimote, Kinects)Mitigation 3: Jeff has several personal Wiimotes available to him for use and there are one

or two extra department-owned Kinects Dr. Daniels has for use. Risk 4: No individuals on the project team have experience programming with

graphics APIs.Mitigation 4: We have built periods of time into our schedule for us to work on developing

the necessary experience with OpenGL and SDL environments to allow us to design and develop software for the Kinect efficiently.

Cost EstimatesItem Cost

Microsoft Kinect (1) for development $150

Wiimotes and Bluetooth adapters $0 (owned by dept.)

Total: $150

Cost EstimatesTask Estimated Hours

# People per Task

Estimated Cost@ $20/hrs

Define Customer Requirements 20 4 $400

Refine Customer Requirements 20 4 $400

Project Plan 40 4 $800

Set Up Website 20 2 $400

Final Design Document 80 4 $1,600

Design Review 8 4 $160

Remove Known Bugs from Wii Wrap 20 1 $400

Develop Installer for Wii Project 20 1 $400

Write Documentation for Wii Project 20 2 $400

Research Software for Kinect Package 100 2 $2,000

Development on Kinect Package 600 4 $12,000

Document Kinect Package 20 2 $400

Testing of Kinect Package 350 4 $7,000

Total 1318 $26,360

Schedule

Hardware - Wii• Wii Remote / Wiimote• 11 buttons + Power• Roll/Pitch/Yaw Gyro• Infrared Sensor• Speaker (unsupported)• Optional Nunchuck (unsupported)

Software Dataflow Wii Lab

Development Software SelectionWii Wrap• Written in C• GUI written in JavaKinect SDK• Written in C++Wii Wrap installer• MSI

How Wii Wrap Works

Wii Wrap GUI

Configure

Set Up Wii Wrap Path

Set Up User Program Path

Run

Run Wii Wrap

Run User Program

Stop / Emergency Stop

Data FlowGUI Window Provides

Output

Wii Wrap GUI

Wii Wrap GUI• Fix Run/Stopped Errors in GUI when the Wii Wrap software

doesn’t execute as planned.• Discover where the lag in display output is (Java Buffers,

Windows I/O)• Integrate GUI into Wii Wrap installer

Hardware - Kinect• RGB Webcam Camera 640x480 @ 30 Hz (center)• IR Projector Projects Grid for Depth Mapping• IR Camera at 640x480 @ 30 Hz collects depth data (right)• Microphones give room audio• Small Panning motor

turns sensor bar from side to side

• Onboard image preprocessing

• Projector power draw requires extra AC power

Thanks to iFixit for some specs data

Kinect - ApproachesNaïve Approach• Develop similar text output program that is based on

OpenKinect, using OpenGL, pthreads, and the Windows port of the Linux USB library functions

• Allow students to generate a spreadsheet of data that they can work on before developing real-time software.

Con: Slow, lots of text processing and possibly storage

Kinect - Approaches

What we need• Provide a GUI environment that would use OpenGL (or an

alternative) to show basic output of the Kinect (depth on video) as well as user modifications to the video data (head replacement, skeleton tracking, etc).

• Want to make use as painless as possible (no pthreads etc to deal with)

Colored image to visualize depth

Kinect - Design• Need to evaluate how to best integrate student code• Concept A : Process-Update Model

• Provide a framework of classes that students modify without needing to know how OO programming works

Process* Update*API for the Student

Called when new frame data is available

Called to update the screen after processing

User Models

* -> function pointers

Kinect• Backend: High-Level Sketch

Acquire Moving Average of Frames

Apply User-Defined Process* Function

Apply User-Defined Update* Function

Apply Changes to Output Screen

• Anticipated User View

Input Screen Output Screen

Test Plan• We will use similar test plans for both the Wii and Kinect.• Initially we will use black box testing methods on individual

modules of the program.• Testing boundary input values such as null, negative values, etc.• Randomizing input values in case edge case is missed

• The second phase of testing will be integration testing.• Testing the entire system as a whole single unit• Making sure all the individual modules function when combined

• GUI testing can be done alongside integration phase of the testing plan.• Consists of making sure GUI is responsive, functions properly

(buttons do the correct actions)

Test Plan (Part 2)• Once the functionality is tested, we will need to do

performance testing especially on the GUI.• Use Eclipse TPTP to test performance.• Test GUI input buffering, load tests, etc.

• Will have Dr. Daniels use the programs to see if they are satisfactory and what changes need to be made.• Make revisions to code based on Daniels’ feedback

Current Project Status• Have been working on developing a software update for Wii

Wrap• Finishing installer for Wii Wrap• Still need to write documentation for Wii Wrap and provide a

distribution setup• Started experimenting with the OpenKinect software and

communicating with OpenKinect founder about release of Microsoft SDK

• Currently preparing a test plan for the Kinect

Work Distribution / Fall ‘11 Plans

• Now – May 6: Complete Wii Wrap Installer (Brenton)• Now – Sept. 1: Identify data stream from Kinect

(Rick/Brenton)• Now – Sept. 1: Make the data streams accessible to

user (Rick/Brenton)• Now – Sept. 1: Developer user engine/API,

investigate graphics (Harsh/Jeff)• Sept. 1 – 7: Document Kinect package (Rick – Lead)• Sept. 7- 21: Testing of Kinect Software (Jeff – Lead)• Sept. 22-Dec: Fix Kinect Software Issues, Complete

Documentation (All)

Questions?


Recommended