We are all familiar with traffic jams in the middle of nowhere. There are no exits, merges, or hills to blame; it's just from the way car spacing is adjusted by drivers. Same thing after an obstruction has been cleared from the road but the traffic jam lasts for another hour as starting up is soon followed by brake lights again.
TACC ought to be a good workaround, at least if there were a dedicated lane for TACC-equipped vehicles. Anyone know the literature on this? What percentage of vehicles equipped with TACC does it take to make a lane flow smoothly? Is anything added by a local-area network to coordinate a string of cars?
There's been a lot of discussion about this, both here and elsewhere, but I haven't seen any actual studies or whitepapers. I've suggested a lane only reachable by AP type cars myself in a couple of them as an alternative to HOV lanes that actually increases a road's car carrying capacity. Train lane is one of the phrases folks were using - swarm another.
You get a fair amount of benefit from just getting the slow human reactions out of the loop - TACC reacts something like five times faster, and notices smaller deviations to start reacting (and can reach 100% braking effort in 125 milliseconds, it if comes to that) - which means it can make smaller corrections safely; the typical rolling stop and go waves come from human drivers being inside their safe reaction zone and thus having to slow faster than the car in front to avoid an accident (which quickly builds from one driver letting off the accelerator slightly through several cars to full on slamming the brakes - all of us delayed because most of us are in a hurry. Ironic, that.)
You can get even better response from cars that are communicating with each other - if the car in front announces to your car that it is slamming on the brakes, you car can start braking within milliseconds of when that car does, instead of waiting to see the speed fall or the brake lights light before starting to brake. That's when you can safely put a string of cars only a few feet apart at freeway speeds and get real density. Swarm logic will also also safe extreme drafting in more normal environments, with decreased energy usage for all cars involved. It may be slightly unnerving for the humans involved at first...
Right now, there doesn't seem to be an established standard for the swarm to communicate, but since Tesla includes Cellular, WiFi, and BlueTooth on ever car and it's firmware updateable for most everything, there's a pretty good chance that the current cars will be updateable to join the swarm once the details get worked out. (Maybe an Ad-Hoc Wi-Fi network among all the cars in a swarm? Your car wouldn't get closer than normal driving distances until is was positive it had solid communications with the car ahead.)
That updateability is one of the unique twists Tesla has put on the situation - they've been able to train the AP programming with millions of miles of driving and human-proctored error detection very quickly because they rolled out the sensors well before the software and have a large and rapidly growing fleet of cars doing it, and they'll be able to quickly deliver a swarm of cars once someone develops the algorithms, provided the swarm software communicates with one of those radios. Lots of incentive for someone to innovate.
Walter
