AI Traffic Lights Could Cut Commute Times

Traffic lights usually coordinate with induction loop sensors or infrared sensors, which are either built into the roads or affixed atop traffic lights to detect the presence of vehicles. When the sensors notice that vehicles have queued, they trigger a change in the signal itself that allows those vehicles to pass. Signals may also be influenced by predetermined schedules, which (if your municipality knows what it’s doing) are based on the flow and needs of traffic around specific times, like rush hour.

These methods aren’t perfect, though—anyone who’s torn their hair out at a long light can tell you that. That’s why Germany’s KI4LSA project, funded by the German Federal Ministry of Transport and Digital Infrastructure, seeks relief in the form of AI technology that can be paired with existing traffic light systems.

The Fraunhofer Institute for Optronics, System Technologies and Image Exploitation, one of KI4LSA’s many partner organizations, helped kick off the project by installing radar sensors and high-resolution cameras at a busy intersection in the city of Lemgo. These devices recorded how many vehicles were required to wait for lights to change, how long they waited, and the average speed at which each vehicle passed through the intersection when its turn arrived. This information was fed into a machine-learning-based computer algorithm, which then generated various traffic light patterns to determine which ones would effectively reduce wait times.

At their most impactful, these AI-powered patterns were capable of improving traffic congestion by 10 to 15 percent—though researchers are still testing whether the algorithm’s predictions would correspond in totality with real-life conditions. Those involved with the project will test out running the algorithm at the Lemgo intersection over the next few months, allowing Fraunhofer to make alterations to the system as needed. Fraunhofer will simultaneously be involved in a similar project called KI4PED, which focuses on pedestrian’s traffic needs and uses the organization’s AI capabilities along with LIDAR sensor data.

Some civilians are skeptical that AI would ever become a facet of American streets, but the odds might be higher than they think. State and municipal governments routinely contract private tech companies to install and manage their public equipment; Redflex and L3 Technologies (now L3Harris) manufacture and operate a wide swath of US red light cameras and traffic sensors, for instance, so it wouldn’t be unreasonable for a company with AI capabilities to swoop in, especially as the technology becomes more normalized.

Continue reading

AMD’s New Radeon RX 6000 Series Is Optimized to Battle Ampere

AMD unveiled its RX 6000 series today. For the first time since it bought ATI in 2006, there will be some specific advantages to running AMD GPUs in AMD platforms.

Time to Update: Google Patches 2 Severe Zero-Day Chrome Vulnerabilities

Unlike the last few zero-days, Google didn't find these security holes itself. Instead, it was tipped by anonymous third-parties, and the problems are severe enough that it hasn't released full details. Suffice it to say, you should stop putting off that update.

This Is a Bad Time to Build a High-End Gaming PC

We're not going to say it's the worst time to build a high-end gaming PC, but if you'll need to get lucky with some orders if you want to pull it off this year.

Hardware Accelerators May Dramatically Improve Robot Response Times

If we want to build better robots, we need them to be faster at planning their own motion. A new research team thinks it's invented a combined hardware/software deployment method that can cut existing latencies in half.