MIT Advanced Vehicle Technology Study (MIT-AVT)
00:00:00.000 | As part of the MIT Autonomous Vehicle Technology Study, we're instrumenting cars with various
00:00:04.920 | degrees of automation.
00:00:06.240 | So let's take a look at one of those cars, a Tesla Model S, and look at our instrumentation.
00:00:12.040 | Inside the car, we have three cameras.
00:00:14.900 | One is looking at the driver's face, and that's capturing things like where the driver is
00:00:19.680 | looking, the drowsiness state of the driver, the emotional state, and also cognitive load.
00:00:25.860 | We have a camera looking at the driver's body, a fisheye lens camera that's capturing the
00:00:31.880 | entire body of the driver, including hands.
00:00:34.600 | That's giving you information about whether the hands are off-wheel, whether the body
00:00:37.940 | is aligned, and further supplementary information about the state of the driver that the face
00:00:43.740 | camera provides.
00:00:45.320 | And finally, there's a forward-facing camera attached to the windshield that's looking
00:00:49.800 | at the forward roadway, and it's capturing everything in the external environment, such
00:00:54.300 | as the vehicles, the lanes, and other characteristics of the road.
00:00:59.300 | Having these three cameras in the car allows us to study driver behavior and interaction
00:01:03.900 | with automation.
00:01:05.220 | So the driver-facing camera, looking at the face, a camera looking at the body, and a
00:01:09.500 | camera looking at the outside environment allows us to understand over hundreds of thousands
00:01:15.280 | of miles of real-world driving, how people interact with these technologies, how we can
00:01:20.300 | have artificial intelligence systems play an important role in keeping us safe and providing
00:01:26.940 | an enjoyable experience in driving.
00:01:29.900 | We have now, to date, collected 275,000 miles of real-world driving and interaction with
00:01:36.420 | autonomous systems in Tesla Model S vehicles, in Land Rover Evoque vehicles, and a Volvo
00:01:41.820 | S90.
00:01:42.820 | But most importantly, once that data is collected, it's just raw pixels.
00:01:47.580 | 3.5 billion video frames of raw pixels.
00:01:51.020 | We're using computer vision, deep learning methods, to convert those pixels into knowledge,
00:01:56.020 | into understanding of what the drivers are actually doing with these systems.
00:01:59.820 | That comes from the face camera, that comes from the body camera, and the forward-facing
00:02:03.980 | camera.
00:02:05.420 | Understanding comes from actually being able to touch every single one of those frames
00:02:09.100 | and convert them into behavior of human beings as they interact with these artificial intelligence
00:02:15.980 | systems.
00:02:16.500 | 3.5 billion frames of real-world driving.
00:02:17.500 | 3.5 billion frames of real-world driving.
00:02:17.500 | 3.5 billion frames of real-world driving.
00:02:18.500 | 3.5 billion frames of real-world driving.
00:02:18.500 | 3.5 billion frames of real-world driving.
00:02:19.500 | 3.5 billion frames of real-world driving.
00:02:19.500 | 3.5 billion frames of real-world driving.
00:02:20.500 | 3.5 billion frames of real-world driving.
00:02:20.500 | 3.5 billion frames of real-world driving.
00:02:21.500 | 3.5 billion frames of real-world driving.
00:02:21.500 | 3.5 billion frames of real-world driving.
00:02:22.500 | 3.5 billion frames of real-world driving.
00:02:22.500 | 3.5 billion frames of real-world driving.
00:02:23.500 | 3.5 billion frames of real-world driving.
00:02:23.500 | 3.5 billion frames of real-world driving.
00:02:24.500 | 3.5 billion frames of real-world driving.
00:02:24.500 | 3.5 billion frames of real-world driving.
00:02:25.500 | 3.5 billion frames of real-world driving.
00:02:25.500 | 3.5 billion frames of real-world driving.
00:02:26.500 | 3.5 billion frames of real-world driving.
00:02:26.500 | 3.5 billion frames of real-world driving.
00:02:27.500 | 3.5 billion frames of real-world driving.
00:02:27.500 | [BLANK_AUDIO]