Quote:
Originally Posted by STK
When the price of cars plummeted with the Model T, the transition to cars was baked in. AV is headed in the same direction. If the safety issue is solved, the transition to AV becomes inevitable. And I guess I should have asked the AV naysayers whether they think the shift from IC to electric is now baked in? What were they saying about that 10 years ago?
|
I'm not sure the Model T is a good analogy for this case. The Model T basically replaced the horse. AV is really just replacing the driver. The automobile was far a better technology, even in its rudimentary stage 110 years ago, to replace the horse as compared to AV replacing humans today.
Regarding solving the safety issue. The engineering challenge is this, develop a sensor suite, communication hardware, control electronics and control mechanisms, and software all to replace the human. These various technologies are all in a pretty advanced state and have been for several decades. Military weapon systems have had such tech since shortly have the Reagan build up of the military. Space exploration drones have been exploring the solar system and planets since the late 1970's. Just two examples.
The autonomous traffic management system either has to be (a) primary individual control at the vehicle level, or (b) primary traffic management at the system control level. Primary individual control is what we have now. It's humans driving cars and reacting to every near-field vehicle in your realm of influence. Situational awareness, experience, and vehicle control skills are massively important to this traffic control methodology. There is some system traffic control in that mix, being road signs and traffic signals. The other methodology is primary system traffic control where there is a surveillance system monitoring vehicle position and each flight (drive) is pre-planned and constantly monitored, this is the methodology the world's air traffic control system uses. There is some low-level human control in the traffic management system, but it is the system that controls the flight (the drive). Situational awareness, experience, and vehicle control skills are important here too, but system redundancy and infallibility are massively important. My informed opinion is that the DOT will pick (b) in a truly Level-5 autonomous traffic environment. Industry thinks it can manage traffic via (a) at the vehicle level to achieve a truly Level 5 traffic environment.
The industry methodology (a) of vehicle-to-vehicle separation relies on situational awareness via those technologies discussed earlier. The engineering challenge is to get those technologies at a level of infallible redundancy produced at an economical price. When someone can point to software that is bug-free operating on a infallible hardware platform that humans have so far produced, then I'll take interest in what they are saying about vehicle-level traffic control autonomy. The air traffic control system is really very good. It's vehicle collision avoidance technology is excellent (try to think to the last time when two commercial airplanes have collided in mid air, or even on the ground), but it is chock full of redundancy and relies on several disparate technologies if one of the others fail. But what makes the air traffic management system safe is the distance kept between aircraft; 5 miles enroute flight AND altitude (ground vehicles lose altitude as a space to place vehicles). And keep in mind there are tens of thousands of humans involved keeping airplanes separated from each other. The air traffic management system (b) probably doesn't scale-up to control automobile traffic in any economic fashion. I don't think the industry methodology of vehicle-to-vehicle management is economically infallible to the level of drastically reducing vehicle collisions. At it's end-state of Level 5 operation, I don't think the rate of vehicle occupancy deaths will reduce much under an AV system, unless traffic density is drastically reduced and vehicle speeds are lowered by at least half. Less people getting around at a slower rate, doesn't seem very progressive.