4/10/19: New project video created! Check it out on the home page.
4/8/19: Active deterrent system and microphones INTEGRATED. The camera now turns to the position of the clap! Next step is to integrate the camera vision features to identify and track the perpetrator (clapper).
4/4/19: Deployment mechanism further improved and camera (with 0-90 degree control) mounted at the end.
3/25/19: 360 degree rotation gear calibrated and precise angle control achieved. Pneumatic cylinder for camera and deterrent deployment was mounted!
3/15/19: Active deterrent system begins to take shaped with main 360 degree gear printed and motor mounted to drive it!
3/7/19: The Intel RealSense and Facial Tracking/Detection have been integrated, birthing a system that can detect and track faces while also taking precise distance measurement. Next step: integrate with gunshot position output from localization subsystem to begin perpetrator detection!
P.S: next, next step: facial recognition (coming soon)
3/3/19: Facial detection and tracking has been vetted! Next step is to integrate this system with the depth information from the Real Sense.
2/27/19: The first revision of the lighting system module has been printed!
2/20/19: The beginnings of the camera vision system was flushed out, with the video feed being filtered by depth. This is the first step in identifying the shooter given an accurate shot position from audio localization system.
2/19/19: Team Interview with Laurie Vasquez about our project and team members. Stay tuned for the article!
2/7/19: OpenCV installed and tested and the Intel RealSense has been ordered to start developing the camera vision system.
2/6/19: Lighting module prototype printed!
1/11/19: The team was informed of its acceptance to the Harris Senior Mentorship Program and are incredibly excited to move forward with more funding and guidance from seasoned engineers at Harris!
11/27/18: Preliminary machine learning results show ~85% accuracy in determining that a gunshot occurred and ~94% that a gunshot has not occurred. Research will continue into better algorithms and perhaps better data collection.
11/25/18: New hardware discrimination circuit accompanied by altered time difference algorithm tested. Phenomenal results observed! (see video below)
11/21/18: 2200+ Audio samples collected to begin 'gunshot' discrimination machine learning. (For practical reasons, the gunshot is supplemented with a clap)
11/15/18: Project video created and submitted to Harris for its Senior Mentorship Program. A short teaser of the full video, showing our current progress in the lab, is linked below!
11/7/18: Small Scale Real-Time Localization Testbed (with GUI) demonstrated!!!!
10/31/18 - HUUUUGE Milestone Update! First real localization (verified) with hardware and microprocessor setup. Furthermore, hand-etching of Microphone signal conditioning boards completed, bringing us one step closer to installation of a fully-embedded system.
10/25/2018 - Double Update! Got the Kinect working and began feasibility investigation
10/25/2018 - Successfully demonstrated working bandpass filter with Arduino Audio capture system!
(Steps represent transitions between known frequencies of 1 kHz, 2 kHz, 3 kHz, 4 kHz, and 5 kHz respectively, and the amplitude of the steps demonstrate proper attenuation/gain)
10/23/2018 - Integrated microphone circuit with Arduino Uno and accomplished three major tasks
10/18/2018 - Successful speed test of Arduino Uno to see if sampling and queue algorithm allow for sufficient sampling rates to be employed in a small scale localization testbed.
10/17/2018 - Successful simulation of gunshot signals with variable time difference, attenuation and noise control. Moreover, extremely successful simulation of our time difference detection circuit (hardware) and time difference algorithm (software)!
10/4/2018 - 3D Localization fine tuned and error improved. First Project Milestone Achieved!
10/2/2018 - Results of the band-pass filter are satisfactory for our current state. The desirable cutoff points and gain were achieved for values we currently assume to be suitable for the system. Values are still open for adjustment after testing is done to make sure that the filter matches perfectly to the desired application.
3D Localization achieved and first simulation output confirmed!
9/27/2018 - Error with respect to sampling frequency and microphone configuration was plotted, providing a tool to select the mic configuration given a pre-defined sampling frequency that satisfies our maximum error requirement.
9/24/2018 - Theoretics of the base equation were explored and it was determined that moving the mics closer increases accuracy. However, moving them closer decreases accuracy due to sampling error. Code was written to account for all error and to find the true optimal configuration for 2D localization.
9/21/2018 - Localization algorithm implemented in code. Error with respect to position in a room for various microphone configurations was plotted and explored.
"L" Configuration vs Line Configuration
9/20/2018 - Theoretical 2D localization algorithm devised!
9/19/2018 - Raspberry Pi set up and first SSH
9/18/2018 - Microphone signal processing circuit designed and tested! Output from one microphone observed.
9/11/2018 - Website created and initial concept articulated in text