I just finished reading an article introducing me to the idea that riding horses and "driving" self-driving cars are similar. I expand on that idea here and introduce practical applications of that knowledge, i.e, a fun way to promote the cars and showing them to be practical.
Basically, one example precedent for self-driving cars is the horse. When horse-back riding, you can either choose to (1) Do nothing, (2) Provide gentle directions (3) Strongly urge a direction. When the rider is doing nothing, the mode of transportation can instinctively go somewhere. When the rider provides directions, the mode of transportation will do them at the earliest convenience, typically immediately. But, the ride won't run into a wall or crash into other rides (a sensible default).
Armed with this knowledge, I think it would be a fun idea to make a video commercial about this. First, the ending to a spectacular action scene, then have a closeup of a cowboy (read: horseback rider) texting and riding. Then zoom out to show all the hard work that the horse has to do to keep the rider safe, like avoiding potholes and other horses. Then, zoom out again to show a parent and child enjoying watching that video inside the self-driving car that is doing the same things.
Random follow-up questions:
- Do we need to hardcode into the car to want to always safe itself from crashing or having a certain threshold or acceleration and deceleration?
- Should the cars think all objects are immovable solids to avoid, or should it use more computer vision (CV) to determine densities (and object types), just in case? Ex: Run into brick wall or bushes?
~ Danial Goodwin ~
No comments:
Post a Comment