Professor Doshi

Peter Petrakis (team manager)
Marcin Celmer
Matt Wilhelm
Tom Stack

When driving down the road we often encounter
traffic lights. The current design of these traffic
lights is that of simple timing, which may cause
drivers to stop unnecessarily. Suppose that
there existed a system that could detect the
density and velocity of the traffic that converged
on an intersection. That the information
gathered would then be used to determine a
new traffic light timing cycle, one that would
cause the driver to wait less at the intersection.
This system will ensure a happier driver and a
smoother flow of traffic.

One prototype will be delivered which will
be scalable from remote control cars up to
actual cars. A deployable field of sensors
will relay information to a central
processing system that will process our
traffic flow and adjust the traffic lights
          Ultrasonic Sensing

●   Sensors will be embedded in existing road
●   Net of sensors located at predetermined
    distance away from the intersection along with
    one sensor at the intersection.
          Ultrasonic Sensing

●   As cars travel over these sensors, the
    signal will be reflected back to the sensor
●   Wireless transmission of sensor net to
    traffic controller
●   Ultrasonic sensors are inexpensive, small,
    and light
●   Implementation time varies from 10ms to
    250ms depending on the range of the
           RF Transmission
●   Master controller has a RF receiver which
    waits for input from the sensor nets.
●   Field bus transmit data collected from field
    bus (sensors) to master controller.
          Signal Modulation

●   Incoming signal to master controller is
●   Signal is filtered and amplified.
●   Result is passed off to DSP for A/D
    conversion and further processing.
       Alternative Detection
●   RFID
●   Infrared
Digital Gateway
        ●   Multiple analog signals
            acquired from individual
            sensor fields.
        ●   Converted to digital
        ●   Processed for desired
        ●   Evaluate the traffic flow /
        ●   Change traffic lights
     Digital Signal Processing

●   Built in A to D conversion
●   Powerful Floating Point capability, free up
    microprocessor to focus on system control
●   Possible ability to determine traffic rank
●   Inexpensive ~ $5
Embedded Operating System
●   Real-Time
●   Low Latency
●   Ability to use higher level languages
●   Less code to debug.
●   Implementation Independent of CPU
●   Possible Architectures including but not limited
    to Atmel, ARM, MIPS, PPC
        Finite State Machine

●   Responsible for maintaining state of
●   Simple/cost effective
●   Verilog
●   Keeps logical errors to a minimum
●   Ensures safe intersection states
●   Uses an Altera chip set
   Traffic Light Controller
We intend to interface with the existing
control system at the intersection while
leaving the existing control system intact.
A strategy could be devised where if the
data collected is not compelling enough
that control of the lights could defer to the
existing timer based control system.
        Demo Intersection

A colored light bulb mock traffic light would
 be used to represent our test intersection
 which would be controlled by our primary
              control system.
       Target Acquisition Tracking

●   Addresses privacy concerns
●   Will not store long term vehicle
●   Size of the memory required is dependent
    on the ID scheme, and length of
    necessary tracking history
             MDR GOALS

•Evaluate the sensor array then using the
collected data the algorithm for the traffic
arbitration will be derived.
•A prototype sensor array will be designed for
the purposes mentioned above.
•The central processing unit will be designed
and the embedded OS will be evaluated.
•The RF transmitter, receiver, and carrier
wave will be designed.

To top