AI Traffic lights
AI is arguably more intelligent than some of the drivers on the road; but it must certainly be more intelligent than most of the traffic light systems that I seem to meet.
Transport for London (TfL) has signed a £6.5m deal for a new control system with outsourcer Sopra Steria. The local government body, in charge of transport in Greater London, handed the contract to Sopra to develop software to crunch a chunkier volume of congestion and road danger information. The new control centre system is …
I heartily agree. Some traffic planner thinks it's a good idea to stop the free flow of traffic at every opportunity. In what way does stopping traffic at crossings at every opportunity help the economy? Why do lights at motorway junctions require stops at two or three lights just to exit or access a slip road? This can only be increasing pollution and journey times. It must be more efficient to allow a long stream of traffic through in turns than to have a slow clockwork like rotation of traffic with multiple stop starts which fails to adapt to loads. Instead of encouraging a smooth driving style this encourages people to race away to try to beat the next light. If AI could learn it should adapt to the condition where no lights are in action and the flow is only slowed when heavy traffic requires it. Far better to limit the speed on a roundabout to 15 or 20mph by installing speed cameras and removing the traffic lights.
I was just thinking, in a gentle fume this weekend how more and more London junctions have what I would term a "no go" settings on their lights. I was stopped at at least two junctions where none of the roads or the pedestrian crossings were green, and I even accounted for the change periods you would expect. I assume this is udner the guise of "traffic management" and will be used to justify some change later on having gathered the right kinds of statistics for vehicle or pedestrian movement.
Biting the hand that feeds IT © 1998–2019