Signal Timing Solution that Does It All
Seamlessly Integrate All Workflow with
One software for a three-step signal timing process
code|GREEN uses a single camera with AI image recognition technology used in autonomous vehicles. It employs Convolutional Neural Networks (CNN) and a deep learning algorithm to process visual data.
The Data Collection
code|GREEN assigns a unique ID to each vehicle and tracks its path through the center of the intersection very precisely. Essentially, code|GREEN fetches turning movement counts directly from the field to the software to enable the creation of data-driven day-of-the-week, time-of-the-day timing plans.
Why Use Artificial Intelligence?
Using Convolutional Neural Networks for Image Recognition. Cadence Embedded Neural Network Summit.
Create Timing Plans
Existing timing plan generation software operates in isolation — one has to physically input data into the software and then print out the output for further application in the field.
code|GREEN solves this problem by directly importing the field-collected data into a single software that uses a mathematical algorithm to generate and deploy a timing plan. Its biggest advantage is the ability to create a wide variety of data-driven, time-specific timing plans, including ones for each day of the week, for special events, off-peak plans with lower cycle lengths, and seasonal plans.
The intelligent algorithm operates by deploying two optimization levels: a Local Optimizer and a Global Optimizer. The Local Optimizer computes optimal cycle lengths and green splits for each movement. The Global Optimization routine creates time tunnels through the signals, enabling vehicle progression.
To create a new timing plan, all one needs to do is to set up an arterial and import collected TMC’s from the centrally stored database. A phase-sequence diagram is then generated suggesting the optimal order and duration of coordinated phases. The software then computes a time-space diagram that illustrates the movement along the arterial and offsets. Manual tweaks allow further refinement if needed. The generated plan can then be reviewed, scheduled, saved, and deployed, all within the software and in a matter of minutes.
code|GREEN is compatible with all types of controllers, cabinets, detector systems, and ATMS software. It is able to seamlessly integrate with and control any type of controller.
The Controller Communication
After a timing plan is generated and approved, code|GREEN can download and operate the timing plan in the field.
The AI processor inputs detector calls for desired phases into the controller that is running in free mode. This does not interfere with vehicle pre-emption or pedestrian operations.