Autonomous drive systems are under development from Mercedes-Benz, Nissan, Tesla, Google, Apple, and others. Today, these systems are nascent but the pace of innovation in this area has never been greater.
Even in this early form, you can find examples of automatic braking systems preventing accidents. Tesla's autopilot has only been out a few weeks, it is beta and still has bugs, but it has avoided accidents that the drivers didn't even see coming. As this technology improves, the cars will gain more abilities.
Here is one example of how things will improve. Today, most cars have antilock brakes. In most situations, antilock brakes are the right way to conduct an emergency stop. However, there are cases when a side skid would stop shorter. However, today's systems value control and maintaining the vehicle in its lane over absolute stopping distance.
Today's simple antilock brake systems have no idea why you are hitting the brakes. They operate the same if you brake because the car in front of you stopped or if a kid runs in front of your car. An autonomous driving system, on the other hand, will have far more information.
Soon autonomous driving will arrive at the place where chess is today. Computers may not be better than the best humans in the world, but they're better than 99.9% of the population. And you don't need a supercomputer to run them; even the chess apps you can run on your smartphone are really good. In a simplified view, autonomous driving is just chess plus physics and computers are really good at both of these.
Autonomous driving is just chess plus physics and computers are really good at both of these.
Professional drivers will take the cars on tracks and preform J-turns, bootleg turns, handbrake turns, drifting, and other get-away stunt driver moves. The telemetry of all of these will be recorded and analyzed. It's unlikely that you'll ever need Dukes of Hazzard or Mach 5 moves to avoid an accident, but it won't be long until autonomous drive systems have complex evasive maneuvers in their bag of tricks.
It won't be long until autonomous drive systems will be able to perform J-turns or drifting maneuvers. In very rare cases, these maneuvers could be employed to avoid accidents.
For example, let's say you are in a future fully autonomous drive vehicle. You are coming up to an intersection and your light is green, so your car enters the intersection. Then suddenly you see to your left, headed directly at you, is a driver coming very fast that is running the red light. Because of a building on the corner, you (and the autonomous drive system) could not see this driver coming. A car is now headed directly at you at a high rate of speed. There is a car in front of you that limits your egress. The autonomous drive system accelerates and then turns hard towards the car. What? Towards the car; is it crazy? This maneuver skids the back end away from the oncoming car. Your car then skids to a stop, sideways in the (unoccupied) crosswalk. Collisions were avoided with both the red light running vehicle and the vehicle in front of you.
If you were in a driving simulator and could try this scenario over and over, you may eventually come to use this same move, but in a real-life situation, it is unlikely that your spur of the moment decisions would be the one that turns toward an oncoming car. Most people would either rear-end the car in front of them or slam on the brakes and prepare for impact.
Kobayashi Maru no win situations are another intriguing aspect of autonomous drive systems, but that will have to wait for another day.
Even in this early form, you can find examples of automatic braking systems preventing accidents. Tesla's autopilot has only been out a few weeks, it is beta and still has bugs, but it has avoided accidents that the drivers didn't even see coming. As this technology improves, the cars will gain more abilities.
Here is one example of how things will improve. Today, most cars have antilock brakes. In most situations, antilock brakes are the right way to conduct an emergency stop. However, there are cases when a side skid would stop shorter. However, today's systems value control and maintaining the vehicle in its lane over absolute stopping distance.
 |
| Professional Driver Drifting Around A Corner |
Soon autonomous driving will arrive at the place where chess is today. Computers may not be better than the best humans in the world, but they're better than 99.9% of the population. And you don't need a supercomputer to run them; even the chess apps you can run on your smartphone are really good. In a simplified view, autonomous driving is just chess plus physics and computers are really good at both of these.
Professional drivers will take the cars on tracks and preform J-turns, bootleg turns, handbrake turns, drifting, and other get-away stunt driver moves. The telemetry of all of these will be recorded and analyzed. It's unlikely that you'll ever need Dukes of Hazzard or Mach 5 moves to avoid an accident, but it won't be long until autonomous drive systems have complex evasive maneuvers in their bag of tricks.
For example, let's say you are in a future fully autonomous drive vehicle. You are coming up to an intersection and your light is green, so your car enters the intersection. Then suddenly you see to your left, headed directly at you, is a driver coming very fast that is running the red light. Because of a building on the corner, you (and the autonomous drive system) could not see this driver coming. A car is now headed directly at you at a high rate of speed. There is a car in front of you that limits your egress. The autonomous drive system accelerates and then turns hard towards the car. What? Towards the car; is it crazy? This maneuver skids the back end away from the oncoming car. Your car then skids to a stop, sideways in the (unoccupied) crosswalk. Collisions were avoided with both the red light running vehicle and the vehicle in front of you.
If you were in a driving simulator and could try this scenario over and over, you may eventually come to use this same move, but in a real-life situation, it is unlikely that your spur of the moment decisions would be the one that turns toward an oncoming car. Most people would either rear-end the car in front of them or slam on the brakes and prepare for impact.
Kobayashi Maru no win situations are another intriguing aspect of autonomous drive systems, but that will have to wait for another day.
0 Comments:
Post a Comment