Quote:
Originally Posted by saifmassoud
This is not autonomous driving, in fact, this is silly. Lets look at facts and projections.
At the moment, we have intelligent "autonomous" driver assistance features at level-2 autonomy. an example would be BMW's Driver Assistance Package, where keep lane-assist, braking and steering is handled by the car in limited situations: clearly marked lanes on the freeway/highway with good weather. And even so, these features are active for few minutes at best, before driver intervention is required.
We are FAR AWAY from truly fully autonomous transportation. Let's look at some realistic timelines:
Mid-2020's some car manufacturers (probably Tesla at first) will offer FULLY AUTONOMOUS driving on Freeways/Highways WITHOUT ANY human intervention at any time and in ANY WEATHER SITUATION. Tesla can do this now but not for all times and not in all weather situations. BMW & Mercedes Benz should reach this level of Autonomy as well (Level 3) by mid-2020's.
By the late 2020's / early 2030's the above will be replicated in NON-Highway-Freeway driving: in city streets, highway exits, suburbs, etc., which is very challenging, therefore, some limited driver intervention might be required.
By the mid 2030's, at this point, we should have FULLY Autonomous transportation. Whether it a ride sharing vehicle or a personal vehicle, the vehicle will be able to pick you up and deliver you to your destination WITHOUT any driver intervention, at all times, at all weather conditions, and in all situations. This is still 1.5 Decades away. So let's be realistic here. So for the time being, lets enjoy the concept of driving.
|
This is the most rational response in this thread to the question, but no one has mentioned the Federal Government (for the USA) has not yet even been involved with the implementation of autonomous transportation. I've stated this in other posts; the Federal Government literally regulates everything to do with transportation and automobiles. The Feds regulate how roads are constructed and operated (lane marking, signage, etc.) down to regulation of how cars are designed and constructed, even to the environmental impact of the paint used on the body. Yet industry and the general public think the US Federal Government is going to let private transportation manufacturers (GM, Audi, Ford, etc. ) and big-data (Google) develop autonomous transportation WITHOUT intervention and regulation by the DOT. It's not going to happen that way.
When the Feds finally figure out how to get involved, you can add 30 years minimum to your 1.5 decades. I work in the air transportation industry as a private contractor in the development of next generation air traffic control. It's taken the FAA nearly 2 decades to close on the envelope of safety to just reduce the separation of en-route aircraft in-half from 5 miles to 2.5 miles. The Feds are not going to have an easy time with adopting autonomous driving.
Highway congestion is not solely the result of poor human driving; it is more the result of road capacity. When one exits the freeway to a secondary road they eventually have to stop at a traffic light controlled intersection. When traffic volume gets heavy there is not enough room on the secondary roads to allow for the free flow of traffic, which backs up traffic onto freeways, which makes merging on and off the freeway difficult, which slows traffic. Autonomous driving will not solve the issue of capacity. The ability of autonomous driving to reduce or eliminate intersection congestion is a serious complex engineering problem to solve, which vehicle-to-vehicle avoidance control technology is not going to be able to adequately control. It will take a "flight" management system to manage individual trips to time-phase traffic to prevent pathway collision avoidance (i.e. prevent two vehicles from being in the same place at the same time). That brings up privacy issues where the Government will now know an individual's flight plan. And all of this depends on if the tech can be developed with enough redundancy to eliminate software and hardware failures and fit inside a vehicle to allow room for the drivetrain and passengers and their luggage.
The cost to make real, true autonomous transportation at even the level of safety currently realized with humans in the loop, will take tens of trillions of dollars. To make ground transportation even safer (deaths per million miles driven) will take even more funding. The level of autonomous driving a Tesla is capable of is infantile compared to what is needed to achieve Level-5; and really, what does Level-5 mean without a Government standard to measure it against.