The next big advance has to be cars that drive themselves (in mixed traffic). See the DARPA Urban Challenge
(1) it increases people's range, because they can sleep, work, etc. in their vehicle.
(2) ultimately, (version 2.0) we can put children and other mobility impaired into the vehicles, and send them on their way.
(3) the car can then park itself. (providing door-to-door service, reducing access/egress time for users in cities and saving on parking costs)
(4) it can go faster as the computer has faster reflexes, though it is still limited by braking speeds.
(5) it can close gaps and therefore increase capacity slightly (depending on how mixed the traffic is).
(6) it is deployable now (assuming it works) as it requires no new infrastructure. The requirement for both new vehicles and new infrastructure (the chicken and egg problem) is what has befallen most previous next new things in transportation (think Personal Rapid Transit).
I believe many of these vehicles will in general be smaller (think Bill Garrison's work), maybe 2 passenger, but perhaps configurable so that an attached platoon can save energy through aerodynamics and space for parking. Say you can chain a few of them together at home, for instance, then a family would go out, but if not everyone were going, the vehicle would be right-sized for the group, with only a little bit of slack.