Eye Imaging Technology for AI

Robotics has come a long way in just a short span of time, engineers working in the field of robotics are always going to come across hurdles hard to overcome. One of these hurdles is vision. Robots cannot have eyes like humans with retinas which allow us to see and interact with the world.

The answer though may rest in optical coherence tomography (OCT) one of the imaging technologies that robotics companies are integrating is the LiDAR (Light Detection and Ranging). This form of vision for robotics is generating lots of attention and investment from car manufacturing companies throughout the world.

This type of vision works just like radar but rather than radio waves being emitted LiDAR uses beams of light. These are emitted and then bounced back to the sensor, the time these light beams take to bounce back is then calculated using the velocity of light (Speed of light) which is calculated at 299,792,458 meters per second. With these calculations AI can build up a rendering of the objects in its optical path.

For a computer system to build up a rendering of complex areas and moving objects this requires large amounts of calculations and time. For self driving cars this is a massive draw back as times is of the essence.

Researchers are focusing on Frequency-Modulated Continuous-Wave Radar (FMCW Radar) LiDAR. With FMCW a light beam is sent out that can continually shift between frequencies. This overcomes the disadvantages standard LiDAR detection systems which cannot determine target range due to the lack of a timing mark.

With the ability of distinguishing between specific light frequencies allows FMCW LiDAR systems to work in any light conditions and at high speeds. These systems have a very high accuracy of range measurement, the ability to measure a very small range, or distance to the target and the ability to measure the target range and the velocity simultaneously.