1. Autonomous Brake Light Communication Phil Osteen Robot Name: Traffic Flobot EML 5666C, IMDL November 25, 2008

2. Outline Motivation –Practical theory behind the project –Ideal implementation Project Description –Impractical implementation of theory –Logical Diagram, description –Sensor package Conclusion

3. Motivation Dangerous high traffic driving scenarios –High Speed driving, Driving at night, Driving on wet road –Depending on driver, brake lights mean A) Get ready to slow down B) Stop! Possible Solution –Automate “brake light” system to detect brake lights ahead of vehicle and apply brake lights –Basic form of vehicle communication

4. Traffic Flobot Project Drives autonomously until object is detected –Classifies object as a vehicle or an obstacle –Driving vehicle emits blue light –Stopping vehicle emits red brake light If lead vehicle is stopping, robot waits behind it –If lead starts driving again, robot follows –If not, timeout causes passing behavior

6. Moving/Stopping LED Array

7. Component Setup –LCD –3 Servo Motors –CMU Cam –2 Sonars –2 Infrared detectors –Infrared emitters/Red LED Array (not shown)

8. Sensor Package Lead vehicle –Red LED brake light array –Blue LED drive light array Follow vehicle –2 synchronized LV-MaxSonar EZ1 sonars Obstacle Detection/Avoidance –CMU camera Vehicle recognition Stopping, moving classification

9. Conclusion Vehicle communication can be achieved using LED arrays and cameras Camera detection and following is possible in controlled environment –But not robust in natural environments Intelligent and functional robot cannot be built for less than $150

10. Future Work Use other sensor for brake light detection –IR detector possibly less sensitive to environment Use DC motors to drive vehicle –Servos work, but are slow Try to apply practical theory in a more practical way –Test outside, on a larger scale

11. Questions? http://plaza.ufl.edu/prosteen/IMDL/IMDL.htm