Connected Signals provides a comprehensive traffic signal data infrastructure.
Connected Signals’ prediction engine uses advanced machine learning and statistical techniques to integrate a broad range of information, including current light state, light timing plans, time of day, vehicle and pedestrian calls, and historical behavior to determine when a light will change, and certainty about the prediction. In some cases, variations in timing plans, emergency vehicle preemptions, or other factors limit prediction accuracy. The predictive information is made available to applications and each application can then determine whether the prediction is strong enough for its intended purpose.
Some applications, like an exact red light countdown, require a great deal of predictive accuracy. Others, such as vehicle powertrain management and “green wave” management, which determines the speed necessary to catch the green phase on a series of lights, can be more forgiving of limited amounts of uncertainty. For applications like these, prediction windows and confidence information can be provided.
Cell phone GPS systems are typically noisy and relatively inaccurate, making it difficult to determine exactly where a user is and how fast they are going. This can be an issue when trying to determine which light a user is approaching. It presents even greater problems when trying to detect when a red light is being (or is about to be) run, where false positives could be distracting.
We have developed advanced techniques for localization that help address these issues. Some of these are already fielded in existing applications, while others will be deployed in future products. For vehicle-based applications, even more accuracy can be achieved by obtaining information from vehicle systems (GPS, speedometer, turn signals, brake pedal status, etc.), resulting in a broader range of supported functionality.
To provide accurate traffic light information, you need to know:
Other information, such as the location of stop signs, can also be useful. Unfortunately, different jurisdictions maintain different levels of records concerning each of these. For example, in some cities the mapping of phases to traffic movements can only be determined by physically opening each light cabinet and observing how that light was configured. In other cities, details of timing plans may exist only in the traffic engineers’ notebooks.
Connected Signals has handled the differing amount and reliability of information available in different cases by developing a wide range of automated and semi-automated tools that refine the available information. These range from automatic detection of errors in city maps to determination of phase maps and timing plans based on analysis of traffic flowing through intersections. These tools make it possible for us to bring cities online quickly, accurately, and with reasonable effort, even when not all desired data is available.
Connected Signals offers a free hardware solution to municipalities wishing to provide their data to us. V2If is a small device, about the size of a pack of cards, that allows municipalities to securely and easily distribute real-time traffic signal data to selected third parties of their choosing. V2If uses two Ethernet ports to preserve network security. One port plugs into the switch that connects cities to their traffic lights, while the other connects to the outside world. As a municipality, you retain control of your data and its distribution.