Coordinated traffic signals that adapt to traffic flow
There's an accident here on I-5 south and cars flock to side streets to avoid the delays. One lane of the interstate is blocked and miles and miles of backup occurs. But then, the side streets become congested too, then the side side streets etc.
The problem is that the traffic signals on the side streets do not adapt to the radically changed traffic pattern. Sure, some signals are programmed by the hour of the day; others are set not to change until a car arrives on the "side" street. But what about when there's an accident? Sometimes you see a cop out there directing traffic through the lights just to get things going.
I believe that the traffic lights can adapt to the traffic by two simple additions to the local control systems. One implements a car counter to detect traffic (backup and flow rate, etc). The other implements a multi-light synchronization so the entire system can adapt in coordination. It's really a simple concept - maximize the flow of cars through an intersection and compensate for nearby congestion by knowing which intersections will be receiving exiting traffic and which will be providing incoming traffic. (Optimal flow isn't local to one intersection only yet can be reasonably achieved within a small radius of nearby intersections, cooperating with each other.)
The car counter technology could be: radar based units, "photo camera" add-ons, feedback from pressure sensitive pads in the road, those tubes they put down to monitor traffic, or even car based beacons (eventually). If you have a cell phone today in your car, you have a car based beacon already. The interconnection of traffic signals / systems could be implemented easily with secure, short burst digital wireless technology. The control algorithm would be a blast to write.
Basically, in my drive home, when the freeway is clogged and people exit to bypass the accident via side-streets, the intersections that are passed through in the bypass route are adjusted (even slightly) to allow more traffic in the most travelled direction. The system and the traffic would seek equilibrium - a balance in traffic flow throughout the whole network of roads. After the congestion dissipates, the system could return to "best guess" settings derived over long term pattern analysis.
This wouldn't have to be done everywhere - but instead could be applied in areas of high "rush hour" congestion and more-than-typical re-routings.
Reward: I just wanna get home in a reasonably predictable time.