Physics and Autonomous Driving

How can Laws of Physics be used to make Autonomous Cars Safer?

Physics is Real and Same all over the World.

In Physics and Real World, no two objects can be in the same space at the same time.

We use this law of physics to make our vehicles safer.

3-D world in 2-D photos or videos

Our world is 3-D and we invented cameras to capture a snapshot of it into 2-D.

In Autonomous Driving, one of major approaches is to use multiple camera to capture images from the environment and "re-stitch" into the 3-D world. This approach is based on Stereoscopy.

Stereoscopy uses multiple 2-D images to create a 3-D scenes from stitching a slightly different view from individual 2-D camera. An example is our own eyes that are placed slightly apart, and how our left and right eye images combine to give us depth perception.

Camera based approaches are error-prone especially at real-time speed that vehicles run at and it is why we have example of "lane-ending", full trucks not being detected in Autonomous Vehicles as well as completely not stopping for even children (Video Link to The Independent Newspaper).

Alternative to Camera-only or Sensor-only Approaches

In our increasingly complex environment, one approach that approaches the accuracy and safety of LIDAR-based sensors with the ubiquity and affordability of Cameras is to use a combination of a sensor with Camera. For most cases, Camera does a great job especially when the margin of safety is large but to be really safe, a LIDAR or Laser based solution is optimal.

Conclusion

Safety can't be a second-thought, it has to be designed into the system from the beginning.

We propose to use laws of physics, where no two objects can be in the same space at the same time to enhance the experience of our Autonomous Vehicles.