Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

452 Proposal Presentation

No description
by

Paul Schroeder

on 1 December 2012

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of 452 Proposal Presentation

Acoustic Gunshot Detection Rebekah Bartlett
Paul Schroeder
Andrew Ou
Baiyu Wang
Whitney Witherspoon Sponsored by:
Harris Corporation Wireless Link Considerations Range
Multipath
Line of Sight
Different ranges to different transmitters
Time delays Wireless Link Considerations Simulation determines minimum requirements
Hardware and software final selection will be based on requirements RF Hardware Considerations Meet system requirements as defined by simulation
Straightforward to assemble/set up
Focus of project is on detection and localization
Links must be fully operational before system test can begin
Meet budget requirements
Interface with microcontrollers and FPGA CC1100 Evaluation Module from TI
Available in 433MHz and 916MHz bands
Programmable to several modulations and channels
Comes with antenna and front end
Linx RF modules and antennas
Will need FEC and synchronization performed on microcontroller
Requires more assembly, but lower cost RF Hardware Options Currently considering ~433MHz and ~916MHz due to larger range relative to 2.4GHz
FDMA system due to robustness to dynamic range and interference
Separate antennas/RF modules on receive end (multichannel receivers are costly pre-made and nontrivial to design and build)
Possible Block Diagrams follow: Current System Design Block Diagram, TI CC1101 Block Diagram, TI CC1101 Block Diagram, TI CC1101 Linx RF Modules Linx RF Modules Agenda Overview
Detection Model
Communication Model
Localization Model
Hardware/Concerns
Questions/Comments Gunshot fired in urban environment
System to detect and localize the gunshot
Communication system to relay information
Accurately filter gunshot from similar sounds Project Overview 1. 3 x C5515 DSP for detection
2. 3 x CC1101 Pair RX/TX
3. 3 x ADMP Microphone w/ Breakout Board
4. Altera DE-2 Board Hardware List Project Concerns Localization: Multilateration Gunshot Event Mic C Mic B Mic A Microphone array placed at known locations
Record Time Difference of Arrival (TDoA)
Calculating distance difference, solve for source location Localization: Multilateration Robust detection
Realistic detection training
RF multipath and range
Timestamp accuracy
Hardware integration
Demonstration http://proceedings.esri.com/library/userconf/proc99/proceed/papers/pap586/p586.htm Milestone 1 November 8th
Functional and accurate simulation model (Matlab)
Milestone 2 November 27th
Full Hardware Implementation
Reach Goals
Firearm identification
3D projection
Advanced information display Project Timeline Machine Learning
Artificial Neural Networks
Supervised Learning
Established technique for detecting gunshots A More Advanced Approach Detection Algorithm Train the Neural Network offline (using Matlab)
Transfer classifier to each DSP board Summary Corpus
A set of samples, each being labeled as positive or negative match
The corpus is fed into the neural network and through an iterative process the weights are optimized. Training Confounding sounds
Fireworks
Back-firing Cars
Environmental Effects
Reverberations
Low-Pass effects Problem with basic approach Look for
Impulses
Spectra that are similar to a sample gunshot
Short times
Check if a set of samples meet criteria A Naïve Approach Image: http://www.dspguide.com/ch26/2.htm Fully interconnected input samples
Each connection is weighted
Training process selects weights and develops relationships between samples
Weights are incomprehensible but work Artificial Neural Network Questions? Hi
Full transcript