Uncertainty in Robotics

Using the word certain the way we'll use it in this Nanodegree, nothing in the previous question is ever certain. Let me explain.

1. What other traffic will do: People are impossible to predict with certainty!
2. Where you are: It may seem like you know where you are when you drive, but you don't. At least not with complete certainty. You may know where you are with sufficient certainty, but if I asked you how many millimeters away from the center lane you were, you wouldn't know.
3. How fast you're going: The same reasoning applies to knowing your speed. You can get a good idea of how fast you're going by looking at your speedometer (which measures your speed), but these measurements are never perfect.
4. What will happen when you turn the wheel: A car is an imperfect mechanical system. If you turned the steering wheel by the same amount 100 times, the car would turn a slightly different amount every time.

As humans we solve these problems in a variety of ways. Number 2 and 3 we solve by saying "ehh, I actually don't need to know exactly where I am or how fast I'm going, I just need to know those quantities with a high degree of certainty". Number 4 we solve by using our brain as a high performance adaptive controller. And number 1… who knows how we handle number 1…? (just kidding, you'll learn more about this in the machine learning course at the end of this curriculum).