You're forgetting a very important factor: the automatic transmission can only respond to the drivers actions. Unlike the driver it has no knowledge of future plans. It will happily shift up 2 seconds before I hit the brakes because it has no idea I'm planning to take a left turn in 20 meters.
> It will happily shift up 2 seconds before I hit the brakes because
> it has no idea I'm planning to take a left turn in 20 meters.
Purely hypothetically, a car could be fitted with an operator control to convey that sort of intent. It could be something the driver could operate without taking their hands off the wheel — perhaps a stick on the left of the steering column?Surely the same data that is used for automatic braking, lane assist etc could be used to improve timings on automatic gear shifts?
I don't think that data is very useful in city driving where most of the shifting happens. No radar or lane detector can detect my intention of taking the next exit on the left.
That being said, I remember reading a piece a while back about a car manufacturer working on linking the car's navigation/GPS to the automatic transmission. So as long as you enter the destination in the GPS it can make more informed decisions.
You making that claim based on how automatic transmissions currently work, right? And not on every way they could work.
Purely from logic, your claim has to be false: Imagine a humanoid robot sitting in your driver's seat, driving your manual transmission car, its robot arm doing the shifting exactly as a human would. Your manual car has been turned into an automatic and it is not less efficient.
I suppose you could argue that the robot, being part of the car, has an energy requirement. I would then argue that we could shrink the humanoid robot down to the size of chip with sensor inputs, and replace the robot arm and gear shift with a solenoid or whatever, thereby saving considerable energy.
By this thought experiment alone, I don't see how anyone can make a statement about automatics always being less efficient.