It has occurred to me that the simple specification for smart car reliability is "that it should not fail in any way at any time under any conditions". Until that condition is available we will need to have drivers paying attention and able to take control.
Driving is one of the daily joys I have in this world. In my 20 mile commute I can take any of 15 or so routes and each route has challenges. Each corner has its entry, apex and exit. No matter the speed, you can hit the apex and accelerate from 21 to 24 or in your mind from 75 to 120. Some of my most entertaining conversations are with (or about) other drivers while driving. Before I will tolerate or support autonomous driving automobiles, I would suggest a remote shut-down that would count the number of electronic "complaints" received and once the magic number was exceeded, the automobile would stop. That way we all could vote on the behavior of our fellow drivers (and get voted on ourselves). We cannot abdicate our responsibility to drive and act civily to a piece of technology. No matter what the lawyers try to prove, we are responsible for our actions and we do not need more technology to further separate us from those responsiblities. This rant is in no way relevant to the technological advances demonstrated in an autonomous vehicle, just to the justification thereof.
On my last trip, I was passed by a Mustang, traveling so much faster than the posted 75 mph, that my SUV rocked. The driver cut around traffic on the left shoulder...sporting a space saver spare on the right front.
Just last month, the Exxon pipeline ruptured and spilled over ten thousand gallons of tar sand into an Arkansas neighborhood.
My SUV spent several days in the shop just this past December when I drove through a local rain puddle in the middle of our street (been there for the 28 years I have lived on the street). The ignition computer AND starter shorted out with the minor splash (splash shields were missing...only maintenance ever done on the vehicle was by the dealer)
My point is that the smart vehicle/autonomous roadway system success depends 100% on built in carefully thought out hardware and programs with redundant systems and meticulous maintenance.
The weakest link within any system is people. As a species, we take shortcuts with maintenance and common sense, then appologize profusely when disaster happens.
Commercial aviation has periodic disasters even though it is run by professionals...who are funded by political whims. I doubt that the automotive corrallary will see FAA style dilligence.
If we ever do get any of the smart cars into general use they will undoubtedly be the slowest vehicles on the road, and the most irritating to every other driver. The reason is that they will all be programmed with the lawyer sitting right next to the code writer and insisting that the software must not exercise any judgement but only do what has been proven to be safe. The result will wind up being way too safe, probably to the point of paralysis. That is what will ultimately kill the madness of the self-driving car, since they will never execute judgement but only go by the rules that are programmed into them. So the lawyers will destroy what the engineers build. How else could it possibly wind up?
Bobjengr, I think that you have just described the real solution, which is a system that warns us of the dangers. THAT could be very helpful. If a computer system winds up driving the cars, they will of course wind up being safer because they will all be going about 15MPH. So the computerized safe cars will waste a whole lot of time for everybody.
And I agree that way to many people are doing other stuff when they should be paying full attention to driving. If you don't pay attention but still survivr, it doesn't mean that you are good, it means that you are LUCKY.
"80% of all accidents can be tied to driver inattention" - that's not a stat. In fact, most accidents are assigned to speeding and then 'driver error' and then 'agressive driving', but this isn't accurate. Most police officers are not traffic engineers or forensic scientists and, practically, they depend mostly on driver interviews, so the data they report is not all that reliable. The high correlation of traffic accidents with geographic location suggests that none of those things should be at the top of the list. Studies of human performance show that a level of distraction is required for best results - it can't be helped that human brains are messy machines - if all they are allowed to do is think about the driving task for any amount of time, they fail. In a different area, in mundane industrial assembly, the ideal worker is a woman with an arts degree: it's not what they do with their hands but what they do with their brains that makes them able to do consistently good work over long periods of time. A level of distraction is essential to performance; on the other hand, sensory deprivation is not.
Traffic lights have been 'programmed' since day 1. Since then the programming has just moved to software from hardware and integrated networking, traffic sensors, pedestrian call buttons and other accessory functions. The fact is that traffic safety before and after traffic lights does not change except for pedestrian accidents going up. Traffic lights are not a safety measure nor can they improve traffic flow more than a very little. The shape of road systems including complexity and distances bewtween intersections means there is no mathematical solution to the problem so there is no possible on the ground solution either. Smart cars on smart highways may actually help slightly as traffic congestion at near design capacity is the mathematically predictable result of human behavior - traffic jams don't just 'happen', they are mostly a human creation - the advantage of robotics here is not he ability to be smarter than humans but the ability to think differently than humans.
All the same, the smartest safety feature roadways can have is design. This is an area of design that is sadly out of step with respect to the computer age. We always end up with 'dangerous' intersections or killer sections of highway identified by an innordinately high accident rate. Typically, the Rx is more signs, changing light sequences, 'driver awareness', etc. none of which actually has any benefit. When all else fails, the 'cause' is poor driving habits; however, that hypothesis requires that drivers habitually enter a state of stupor when they enter a certain traffic zone and are miraculously better only a few feet further on. Sadly, the worse it gets, the more a declared 'dangerous' traffic zone becomes designed by a committee, primarily composed of amatuers. Engineering departments don't want to admit the design is deficienct. Politicians don't wan't to admit that they weren't willing to pay for the best-practice solution or that their main goal was to satisfy developers. Traffic calming is almost a bellweather of dangerous design, usually after the fact: congestion increases driver conflict which increases accident rates so traffic calming, which consists of adding additional distractions while increasing congestion is implemented as a 'safety' measure. Politicians love it - it's really cheap and it creates the appearance of 'doing something'. What we really need first is streets roads and highways that are intelligently designed.
I commute about 75 miles per day back and forth to work. On Interstate 75 heading towards Knoxville, there is an overhead sign that posts a variety of messages; "watch for bikers", "left lane blocked ahead", "don't drink and drive", "seatbelts save lives"etc. You get the picture. One sign I hate to see is the number of fatalities to date. For the state of Tennessee, as of today, there were 644 fatalities. Horrible numbers. You would not believe what I see every day on the way to work. At least 50% of the drives are using a cell phone. Ladies putting on their makeup. People reading books, maps, newspapers. I actually saw a highway patrolman, one of Tennessee's finest, inserting a magazine into his handgun. I am all for reliable devices that can indicate hazardous situations BEFORE a tragedy occurs. People are NOT going to change their habits but maybe technology can bring about warnings in time to save lives.
The question of whether engineers could have foreseen the shortcut maintenance procedures that led to the crash of American Airlines Flight 191 in 1979 will probably linger for as long as there is an engineering profession.
More than 35 years later, the post-mortem on one of the country’s worst engineering disasters appears to be simple. A contractor asked for a change in an original design. The change was approved by engineers, later resulting in a mammoth structural collapse that killed 114 people and injured 216 more.
If you’re an embedded systems engineer whose analog capabilities are getting a little bit rusty, then you’ll want to take note of an upcoming Design News Continuing Education Center class, “Analog Design for the Digital World,” running Monday, Nov. 17 through Friday, Nov. 21.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.