Self-driving cars are quickly becoming a reality. Now, researchers have come up with a new approach to time-of-flight imaging that increases image resolution of depth perception 1,000-fold. These researchers, called the Camera Culture group, based out of MIT’s Media Lab, have been developing new imaging systems for a decade. This new time-of-flight approach gauges distance by measuring the amount of time it takes light projected out into the environment to bounce back to a sensor.
The MIT scientists recently presented their findings in a new paper published in IEEE Access. This breakthrough technology can give accurate distance readings through fog. Up to this point, that has proven to be one of the greatest obstacles to the implementation of self-driving cars on American roadways.
This new technology has a range of two meters of depth perception. Previously, time-of-flight depth perception systems only had a resolution of about one centimeter, just enough for assisted-parking and collision detection systems on cars presently on the market. An increased range of distance decreases resolution exponentially, but at distances of two meters, the new system has a depth resolution of three micrometers.
A short burst of light is projected out into the environment around a car. A camera then measures the length of time it takes the light to return, indicating the distance of the object that the light bounced off. The further out objects are, the more difficult it is to measure. The length of a light-burst’s accuracy is one of the key factors determining the system’s resolution. Detection rate is another prominent factor. Modern detectors can only make 100 million measurements per second. The light beams are switched on and off by modulators that can blink a billion times per second. It is the detection rate that has been restricting the time-of-flight system to centimeter scale resolution.
Fog has always been a problem for time-of-flight systems because fog has a way of scattering light. Light sent into fog is deflected, and when the light returns to the sensors, it comes back late at unexpected angles. With high-frequency systems, however, the phase shift is larger relative to the frequency of the signal, and scattered light signals that arrive over different paths cancel each other out. This makes it much easier to isolate the true signal, according to preliminary research conducted at Columbia University and the University of Wisconsin.
As technology continues to improve, self-driving cars are getting closer to taking over the roads. Evolving regulations and the law are developing at a slower rate than the technology, leaving many to wonder how accidents and injuries will be handled when cars can drive themselves.
If you or someone you love has been injured in a car accident, you know how hard it can be to get back on your feet. Let the experienced Cherry Hill car accident lawyers at Williams Cedar handle your legal issues so that you can focus on your recovery and begin moving forward again. To schedule a free, confidential consultation, call us today at 856-874-7500, or contact us online.
With offices located in Philadelphia and Haddonfield, New Jersey, we represent people injured in car accidents throughout Pennsylvania, southern New Jersey, and nationwide.