HSV CW

Eat lazer clankers.

I'm just going to briefly talk about them.

So if we use a camera, I use a different model than if we use a range finder.

This is a supplemental system to the localization class I just taught you.

It uses a camera system, together with a laser.

It's a new gaming platform that's been sold at record pace.

This is where machine learning comes into play.

Let me tell you something about the equipment that's on this car that makes it self driving.

We use a lot of laser range finders.

From this we moved onto build Junior, which competed in the DARPA Urban Challenge.

We learned about how maps are being used to localized a moving robot.

But you understand the gist of the solution now, and you've been able to program it.

Here is imagery of our laser system mapping out the terrain ahead.

In our experiments using robots to drive through the desert and through traffic.

Essentially, it's trying to drive on flat ground.

Turning on the spot clockwise costs 10, but turning counterclockwise, which is C CW, costs us 0.

Here is another vehicle, and another one.

You can turn the spindle clockwise or counter clockwise by pushing the CW or CCW buttons to aid in setting tool offsets

Some were related to localization, such as global positioning sensors, and some were related to understanding where obstacles are, like laser range finders.

We quite extensively used laser range finders as an alternative to stereo vision, because they give us very, very good range estimates.

Using radars and lasers, the Google car estimates the distance and the velocity of all these vehicles, and it does so using a Kalman filter.

Here is a 3D model constructed by laser range finders of our neighborhood in Palo Alto, and it's easy to see how 3D points can making amazing 3D models, using techniques like stereo vision or like the laser range finders I just briefly talked about.

What you see here is camera images delivered by the robot superimposed with laser data that doesn't see very far, but the laser is good enough to extract samples of driveable road surface that can then be machine learned and extrapolated into the entire camera image.

Our Google self driving car uses radar on the front bumper that measures the distance to vehicles and also gives a noisy estimate of the velocity.

That enables the robot to use the camera to see driveable terrain all the way to the horizon up to like 200 meters out, enough to drive really, really fast.

The vehicle was based on a Volkswagen Touareg that was equipped with all kinds of sensors like GPS and laser, and it was able to make its own decisions without any human input whatsoever.

Rain, for example, changes the total appearance of the street, and we can adjust this with a single constant factor that adjusts the total brightness of the laser map.

This shows work on machine learning autonomous driving, where we used the laser to identify driveable terrain at a short range and then extrapolate this out into the long range using a machine learning technique applied to computer vision.