In a crowded environment, the further you can see, the faster you can safely go. This applies to everything from flocking behavior to warships.

• High-end sportscars optimize for light-weight and minimum-body-roll. This often means that the most potentially agressive cars (Corvettes, Lamborghinis) have the least visibility. (because they're astonishingly low to the ground)
• Vehicles that have the most distant visibility are 1) semi trucks and SUVs, and 2) any vehicle that changes lanes, during which it's able to look down the long gap between cars, and read the road much further than they otherwise could.
• If one had unlimited money and time, the most obvious design for an "all around vehicle view" would be to mount many cameras on all corners of the car, and integrate them using complicated algorithms, and present one or two unified views that the driver could usefully interpret.
  • on the other hand, the human brain contains one of the more complex and mature image-processing systems around

One possible way to resolve these contradictory requirements is to implement two eyestalks that send an independent image to each eye. This allows the driver to 1) have an extremely lightweight car, while 2) seeing things from as high as is legal, 3) and as far to the sides as is legal, 4) while requiring NO additional image-processing. (though, granted, it requires the driver to wear "fail safe" head-mounted displays)

Moreover, a more complex design allows for movable eyestalks, to accomodate legal limits while extending vision further to the sides when possible.

A motorcycle-based version is even possible, in the shape of a "T". However, it's almost certain to affect vehicle dynamics at least a little, unless extreme attention is paid to light-weight and aerodynamic design.

While the end result seems obviously anthropomorphically-originated (even to infants), in fact it's a natural outgrowth of the above requirements.