210 likes | 352 Views
Dieter Laskowski Jesse Harvey Mark Cataldi. Drowsy Driver Warning System. Outline. Overview Analytical Components Testing Strategy Deliverables Cost Estimates Project Status. Overview.
E N D
Dieter Laskowski Jesse Harvey Mark Cataldi Drowsy Driver Warning System
Outline • Overview • Analytical Components • Testing Strategy • Deliverables • Cost Estimates • Project Status
Overview • Drowsiness slows reaction time, decreases awareness, and impairs judgment, just like drugs or alcohol The driver was apparently drunk and fell asleep when he crashed into the cyclists. One rider was killed and 10 others injured in the incident. The truck driver was driving fast and he was sleepy; consequently, the truck hit a palm tree then it flipped. The day-care bus crashed into a bridge killing the driver and four of the six children aboard. The National Transportation Safety Board said the driver fell asleep.
Overview • One camera will monitor the drivers eyes • A second camera will monitor the roadway • If a situation arises the warning system will be activated, alerting the driver
Cameras • Product: Logitech QuickCam® Pro 9000 • Retail Price: ~$100 • Relevant Features: • Premium autofocus: Images stay razor-sharp, even in close-ups. • Ultra-wide field of view and intelligent face tracking: Keeps you right in the middle of the action. • Modifications: • Infrared capabilities
Warning System • Piezo buzzer • Massage Pad
Warning System Controller • Alert devices are controlled using Freescale HCS12 development board. • Software receives alert commands via serial port on laptop and takes appropriate action: • Piezo buzzer issued appropriate PWM signal. • Massage pad activated using DC signal. • External switch attached to board is capable of inhibiting alert peripherals
Power System • Product: Sima STP-325 • Retail Price: $30.62 • Relevant Features: • Low battery alarm • Two short circuit protected AC outlets • Mounting feet for easy and safe installation in any vehicle to vehicles 12v supply • Cigarette lighter plug provided • Alligator clips and cable for installation • Automatic thermal protection indicator • 325 Watts Continuous, 650 Watts Peak power
Machine Perception Toolbox • Supplies cross-platform libraries for real-time perception primitives, including face detection, eye detection, blink detection and color tracking. • Our Application: Wrapper around Machine Perception Toolbox (MPT) blink detection to determine if user is becoming drowsy or sleeping.
Control System
Testing: MPT Blink Detection • Two team members will drive around together at different times throughout the day to determine how effective the blink detection is under different lighting conditions and real world situations.
Testing: Lane Detection • Two team members will drive around together at different times throughout the day to determine how effective the lane detection is under different lighting conditions and real world situations.
Testing: Warning System • During testing, the inputs will be forced to create a warning condition. When the control system detects this, the alarm system should respond promptly. • Each component will be individually checked to verify that it does in fact turn on when commanded. • The warning system will also be thoroughly tested after integration with MPT and Lane Detection • Peripheral inhibit switch will be tested to ensure that no alerts are issued when the switch is activated.
Testing: Power Systems • When all devices are completed, they will be installed into the car and connected to the power inverter. • The system will be tested with maximum strain on the cars power system
Deliverables: • Drowsy Driver Warning System set up inside of a cardboard mock car • LCD monitor set up outside of the car so the audience will be able to see the results of the Blink and Lane detection. • Car driver will simulate falling asleep to force a response from the warning system • The demonstration will wrap up with a question and answer session as well as an opportunity for volunteers to try out the system.
Project Status: • MPT investigated, primary points of interest outlined • Investigating additional department resources and materials • Initial testing with camera software complete • Lane detection algorithm implementation started