Automated Vehicles Set To Take Over Trucking Industry

In response to something alluded to in another posting today: Don't You Dare Miss That Sign - Article By Old School

Love to hear our new member Jeremy's take on this.

As for automated vehicles taking over the world... Don't hold your breath. From Tesla to Google to Amazon, the reason you won't hear Musk, Bezos, or anyone else talking about a timetable for practical use is exactly because of the unpredictable variables involved with non-interstate operation - many of which are identified by those ever-present signs (blocking out the scenery, breaking my mind!) Don't get me wrong, there are plenty of software scripts out in the wild that utilize pattern recognition and can process/render critical algorithms in just nanoseconds (I mean pretty frightening stuff.) But all of that can be thrown off by the slighest modification to the expected pattern that an important sign might possess (via weather, vandalism, age, etc.) Now, most software can be written to assess and render a high-percentage of likelihood for an equation, so you could have an automated vehicle guess what the sign most-likely is - But who wants to take a gamble like that?

And let's not forget that sometimes there is only a small amount of uniformity in sign regulations once you leave federally funded roads. This can result in various locales even using their own messages, designs, or types when it comes to signs. Especially smaller rural areas where such signs might simply be painted somewhere, etc. Software can use pattern recognition to pick out various words or symbols and interpret their meaning, but that only works when it has something to compare pattern samples against. Variation in local discretion could result in a project that would take years to compile all of the required data for a machine to safely interpret every possible type of statement or warning that a sign might display. And that assumes that every sign was actually in place (not blown over, stolen, vandalized, etc.) when an automated vehicle needed to see and recognize that sign. A person can likely see a twisted, half-bent stop sign after some drunk plowed into it (albeit maybe a little later than optimal) but would an automated vehicle even notice that the sign existed, much less be able to read it? Likely not. And if it could, now just imagine all of the false-positives that could manifest due to that ability.

An argument could be made for electronic signalling devices being added to signs and installed on vehicles (anyone ever use EZ-Pass on the turnpike?) But that would require an unimaginable amount of logistics, including a timetable that might itself take a decade just to design, to say nothing of the fact that somebody has to foot the bill for all that technology and the physical labor required to modify all those signs. And I highly doubt that most muncipality's or even state's would be willing to foot that bill when many can't even address basic existing infrastructure issues, such as hazardous schools and eroding bridges. Well, I guess corporate America would be chomping at the bit to get in on that - Just imagine cruising down the AT&T I-90 "Rethink Possible" Interstate and having to catch the Nike I-77 "Just Do It" exit to head south. (Hey, if McDonalds buys the rights to I-10, does that mean no more Burger King between Florida and California?)

How about the GPS argument? As things stand today, GPS (Navstar?) is just not accurate or robust enough to safely and reliably support a system of this magnitude (to say nothing of all the new variables that arise.) Now, when GPS III finally comes fully online (2025 maybe?) it should be ready for testing, but not full implementation (in regards to safe automated travel.) They have to finish launching all the new sat's first, then align the whole system, and God knows what else... And when GPS III is actually going full steam, everyone is going to have get in line behind aviation, marine, etc. - all the existing civilan services that have already spent years waiting on the release and propagation of that system. Sorry, but GPS just ain't going to get it done for another 20 years minimum.

To be honest, auto-pilot might be great for traversing well established lanes of travel (interstates and highways.) But when it comes to getting off the main routes and navigating both urban and rural routes that have very little uniformity the required technology, the ability to implement that technology (and the finances to pay for all of that) are still decades away. And finally, while we can (and do) make supercomputers capable of astounding feats of technology, those feats are often still limited to static binary decisons (if this, then that.) And the existing software that boasts predictive powers and human-like intuition often falls flat over the simplest nuances and anomalies.

Bottom line: We are far short of seeing any technology that can replace the discretion and decision making powers of the safe and experienced truck driver.

A great article on just how spectacular failures can be due to the simplest of machine miscalculations, check out this arcticle: The Coming Software Apocalypse

The time may very well come when drivers are replaced by technology, but only after a lot more research and development takes place. We may never see that day and I would be okay with that because I don't want a "regular" job and I don't want to see autonomous vehicles taking over the highways and byways. I like driving too much to let a computer have all of the fun.

I can make this even more simple. Last time I checked their are still at least 2 pilots in every multi engine aircraft. They have had auto pilot for decades. I'm not going to worry about something that won't be truly functional until well after I'm retired.