A platoon of driverless trucks equipped with Cooperative Adaptive Cruise Control (CACC) appeared on the 110 freeway in San Pedro. The system controls speed, braking, and spacing of the convoy. The goal is to increase safety and fuel economy.
Three big rigs barrelled up and down the 110 Freeway on Wednesday, mirroring one another in a tight pattern. Two of the Volvo big rigs bore special antennas to “talk” to one another and radar that can detect movement around them. They were accelerating and navigating without human help.
It was only a test, but the partially automated trucks provided a peek into the future of long-haul trucking. The demonstration’s sponsors hope it provides a step toward completely automated transport in the years ahead.
“It’s smooth, safe and efficient,” said Carrie Brown, Caltrans’ district director for Los Angeles and Ventura counties.
Using what’s called “cooperative adaptive cruise control,” the heavy trucks drive tightly together, responding to one another and their surroundings with computerized sensors, saving fuel and releasing fewer emissions. Well-plotted trips would also ease congestion, experts believe.
“Simply put,” she said, “the trucks are driving as a stable unit.”
The demo team acknowledged there are kinks to work out. For instance, the communication system needs to be protected from interference from outside systems, so the security needs to be shored up.
At one point in the test a driver had to take over, but this is an excellent start.
It will not take more than a few years to work out any problems. Click on the link at the top for a video.
Mike “Mish” Shedlock
Once again, for long haul, we already have railroads that can be automated. Even today it not much manpower is needed to take a 100 car train across the country. Trucks will never achieve the fuel efficiency of trains, and trains can be electrified easily. Great Northern and Milwaukee had electrified routes through the Rockies decades ago.
true trains achieve 1,000 ton/miles per gallon and trucks less than 100.
but trucks still have a 50% market share of freight in the US.
On the topic of this test-
This appears to be a test of using (near) SD trucks on a dedicated route from Port of LA-LB to inland warehouses. This is a perfect early adopter niche for SD trucks.
I believe Mish is wrong in his timing. It will not take ‘years’ to work out the kinks in this type of SD trucking. Rather, it will take months. I predict by 2018; there will be regular commercial SD trucks being used on dedicated routes. We could call this widespread beta testing.
By 2019, SD long haul trucking will be quickly adopted wherever dedicated routes exist running from a warehouse to warehouse located close to interstate onramps. A huge amount of trucking is simply trucks doing the same route on a interstate over & over again. Perfect opportunity for level 4 machines. And we can expect a few thousand of these machines by 2019.
By 2020 we are likely to see level 5 beta testing of trucks
By 2021 there will be the earliest mass dislocation of trucking jobs as the big guys quickly transition to SD level
4 and level 5 machines.
Does anyone really think Amazon, UPS, DHL, or JB Hunt is going to resist SD technology ?
Don’t forget Ali Express and Ali Baba.
They can use this technology to drive right past Amazon direct to your door.
A least until Amazon lobbies for Made in America-ish legislation to stop China from undermining Amazons strategically important business of re-selling Chinese crap to Americans.
Trucks have 50% market share because the government built the interstate highway system that enabled them to transport goods faster than the railroads, combined with cheap fuel and capital expenditures for rolling stock that are much less than for railroads. Fuel costs and availability will eventually drive freight back to railroads and water transport.
Government built the freeway infrastructure to transport people. Trucks just kind of piggybacked on it. Even absent trucks, we would “need” an interstate system.
Fuel costs aside, pollution will become a big motivation to move transport, goods and people, onto rail. As will increased expectation for travel speeds. With engines running cleaner and cleaner, the particulates created at the tire/road interface, is accounting for an ever larger share of total pollution. Steel on steel, and maglev, is much cleaner. Travel speeds can also be raised much higher safely on the two latter infrastructures. High enough, that you’ll also make sizable cuts in the nasty pollution produced by jet engines.
Also, with electrification of the fleet supposedly the future, you’ll need an infrastructure able to supply continuous juice to individual cars anyway. Dragging heavy, expensive, toxic batteries everywhere, just to burn through 90% of their juice during the 10% of the travel time that is spent on highways, is too stupid to remain sustainable forever. Smaller batteries will be supplied for local hops. Those batteries will be automatically recharged during freeway sections. With the cost for the power used, billed to the owner of each vehicle.
Technology is now at a level where the kind of flexible and dynamic derailing required to move individual cars onto rail, is at least conceptually attainable.
In addition, the vastly limited event space a vehicle can be expected to face on a controlled access rail network, is much better suited for autorouting, than the current, free for all, infrastructure, Vehicle to vehicle communication, traffic management etc., can be built in from the get go. Reliably interference free. All the other drivers will be certified computers as well, or they’ll be denied access. So no worries about aerodynamically optimal tailgating at Shinkansen speeds running into trouble, just because grandpa mistiook reverse for forward on a farm tractor and then had a stroke, so he’s coming at you backwards down the highway.
What is “SD”?
SD = Self Driving
Self Driving
Apparently automating the trucking industry is easier than introducing new technology to the railroad industry.
Remember that rails were the first nationally regulated industry, and one of the first crony capitalist industries, and one that has a heavily unionized workforce. Unions put the kibosh to technological advancement after they saw their numbers reduced after the introduction of air brakes, then later the diesel electric engine. The Amtrak engineer spends most of their time pushing a deadman button just to prove they are still awake. The workforce of the rails is about as low as they can get now because the unions will walk out if they try true automation.
Highly likely this will assist rather than take-over the task of driving. We already have auto-pilot on aircraft yet pilots are still present. How many deaths have occurred as the result of a truck driver falling asleep at the wheel?
Total takeover is coming and obvious
Cars totally replaced horses
The same construct coming
There may be a period where there is a driver assist, but it will vanish
Unions, Mish. Nobody understands their destructive (political) power better than you. Your model of how this plays out would be correct in a world w/o Democrat Party, right-brain thinking. However, Democrats will always interfere in markets to protect one of their politically-favored constituents, and they will do so in the unionized trucking industry.
Markets >> Unions.
You don’t adopt SD trucks, you pay a lot more, become uncompetitive, fail and go broke,
Conscience — ‘exactly.
Horses had drivers
Cars have drivers
…so….um….err?
Full -100% driver-less – Never. Gonna. Happen.
Now if they build, or dedicate specific roadways that would might be a plan.
Meanwhile the trucking industry is in one of it’s largest driver droughts in history — HELP WANTED — or maybe they are reading this blog and getting out before the hammer falls.
Full 100% driverless will be here in a limited way by 2020 in some places
I agree that driver assist is probably the outcome in the short term of ten years or more for cars, faster for trucks. Many older drivers simply will have nothing to do with it but younger drivers will accept it more readily.
Great progress. The government and the public has neglected rails for many years. Now we are investing billions in sophisticated tech to have the trains run on our roads. First trucks for freight and then cars for passengers. The 1860’s are right around the corner.
Driverless vehicles on public highways is a great example of an economist’s wet dream. These trucks are far from “driverless”, rather there are computer systems that seek to accomplish limited assistance to human drivers who must be alert at all times, like all other drivers.
And running semis together “in a tight pattern” must mean they are run in a line with limited space between them. That may sound good to an amateur, but wait until a car-driver decides to cut into the line to get to an off-ramp – then you have chaos. Or there is an emergency that creates an unavoidable accident – 3 loaded fuel tankers going up in flames rather than one is a real possibility. And if the group has to change lanes quickly, the computer will try to keep the group in line – then you have an extra-long snake trying to find a large enough hole in traffic to move into.
All these “driverless” vehicle systems can do, so far, is keep a vehicle in it’s lane (if the lane is clearly delineated by means that the system recognizes), maintain spacing between vehicles in normal circumstances, and regulate vehicle speed. If the system depends on painted lines on the road, it is useless on a snow-covered road (as was discovered in Scandinavia recently when an automated car drove off a snow-covered road).
What these systems cannot do is read road signs such as “Right Lane Closed Ahead”, “Construction Speed Limit 30”, “Accident Scene Ahead”, “Wide Load”, “Road Flooded”, “Bridge Out”, “Limited Visibility”, etc. And picking out a red light among numerous lights in a typical setting is proving a challenge beyond the abilities of current systems, witness the Uber driverless experiment that was cancelled in California after a car ran a red light.
Converting to driverless trucks would require a total change to new trucks and trailers, along with shipping and receiving facilities, at enormous expense. And the truck operating systems will have to become far more sophisticated and reliable than current systems.
Converting to driverless trucks would require a total change to new trucks and trailers, along with shipping and receiving facilities, at enormous expense. And the truck operating systems will have to become far more sophisticated and reliable than current systems.
Good grief – please think – better yet read
I have addressed all of this dozens of times
For starters, one simple approach is hub-to-hub driverless where long-haul interstate driving vanishes first along with millions of jobs.
Second, the cost of the systems will be far less than the cost of a driver.
These arguments are absurd. Progress is happening at a pace that exceeds my expectations.
Mish I see so thrilled millions of Americans job losses are just around the corner.
Just finished a trip from Denver to Reno on I-80. I’m wondering about how an automated vehicle convoy would manage those construction zones in which everyone has to get into one lane for 5-10 miles. Perhaps the system could be set up so the first truck has a driver to handle the unpredictable events and the rest of the fleet just tags along. When the “right lane is closed ahead”, everyone will be limited to the speed of the retired guy in his under powered motor home and if he winds up in the middle of the convoy, the lead dogs will necessarily have to slow down as well to avoid becoming separated.
Construction closures will generally be known
Sudden closures – accidents – theoretically more problematic, but we have many years to work out a solution.
I am confident it will not be that hard.
This is progress? Eliminating truck driving jobs so that more people go unemployed? This techno advancement for the sake of financial efficiencies, no matter the human result, is just plain stupid. Oh wait – less pollution makes it OK.
The future’s coming. It can’t be stopped.
Replacing humans, at the wheel, reduces driver error and will save lives.
40,000 auto-related fatalities in the US in 2016. Yes, reducing that number will be progress.
Zero Hedge reports a shortage of truck drivers….problem solved.
http://imgc-cn.artprintimages.com/images/P-473-488-90/60/6060/EK4D100Z/posters/j-b-handelsman-the-lemmings-of-montauk-new-yorker-cartoon.jpg
The loss of jobs is always the argument against innovation and increased productivity. The argument has never borne out. Its economic evolution and the jobs lost would be replaced by higher skilled jobs as the workforce readjusts, retrains and adapts. At least in theory. BTW we are already at full employment and have a generation of snowflakes that will not and cannot do any “hard” work. I blame the boomer parents for not instilling a work/responsibility ethics in their special little privileged snowflakes. My kids think I am to much of a hard ass but I explain to them its not personal just business,
How many boomers could program a VCR? (still waiting to meet the first one, even if VCR’s are outdated)
How many snowflakes think configuring WiFi and cell phones is child’s play? (they have been doing it since they were children)
Now what happens when these bored, debt enslaved children realize they can easily hack one of these truck convoys and drive it through the bursar’s office of the university that enslaved them with political indoctrination, without teaching them how to even pay their student loans back?
What happens when the political activist snowflakes get a few bucks from George Soros to drive a truck convoy through the opposition’s political rally? Or through planned parenthood? Or a chemical plant? Or a major sporting event, where steroid injected neanderthals get paid millions while actual students get buried in debt?
Let me guess. The Titanic was unsinkable, and these trucks are unhackable….
I would guess it will be an arms race between security companies building better hack protection vs terrorists or disillusioned snowflakes. Maybe they need to bring back the draft so the useless flakes will get their shit together. The only reason the flakes can’t pay their college debts back is because they believe their degrees in art history or gender inequality will payoff. They borrow money to go away to school and have a “living experience” and are shocked just shocked when reality hits, Of course its everyone else’s fault that these idiots can’t make it.
Not saying the snowflakes are right or wrong — only that they are disgruntled and deep in debt.
I do blame academics for selling magic snake oil, which is the nicest thing I can say about the crap that gets a college degree now. Most of it is political indoctrination, with staggering amounts diverted to administration, assistant deans of everything, fundraising staff, and coaches that are paid millions for “non profit student athletics”. When other entities do the same thing, its called fraud.
Agree that gender inequality degrees are less than worthless, but that is what the schools are selling. Not sure the so-called business degrees are worth much either. Yes, in concept business degree might get you a job… but when you factor in lots of silly “core” courses in political correctness, today’s business degree is not what older generations remember.
Anyway, the disgruntled snowflakes are still deep in debt for a fraudulent product. And they still know more about wifi setup than the older crowd knew about VCRs. If they are now told their worthless degrees can’t even get them a job driving a truck…
You are right colleges offer many degrees that are unproductive from a cost/benefit analysis, They promote liberal ideology on top of it. I find it disgusting to have to pay for that. They make it impossible to graduate in 4 years in order to make more money. They are lengthening the education requirements for some degrees for no good purpose other than to extort more money. When I went to college in the 80’s a physical therapy degree was 5 years, same with pharmacy. Now they are 8 years requiring a PhD. What a crock of shit. College costs have exceeded the CPI more than any other cost. It is insane and you are correct the money has poured into non academic areas benefiting the administration and not the students. I just recently heard that if you took the cost for a year in Harvard at the 1970 baseline price and applied the normal CPI to it, tuition would be about $15000/year vs the >$50k it is now. I hope the free market finds a cost effective solution for a degree like coursera, the khan academy, edx, etc to put those useless corrupt sob’s out of business. If something can’t keep going it eventually will implode. The sooner the better. I must add however, most parent are afraid to hurt their children feelings and make them go to community college first and even to a local university where they can commute to, My second child went to community college for 2 years because he was undecided on what to pursue. We did this per the advice of his high school counselor, who suggested that advice would be good for everyone decided or not. I was upset when I realized how inexpensive community college was and regretted not making my first child attend also, An expensive learning experience In the mean time many hight paying skilled trade jobs/apprenticeships go unfilled because the snowflakes don’t want them..
PS … so far, the disgruntled hackers who dropped out of school (by choice or by necessity) have run circles around security “experts”
We have never before been at a point where AI could replace brains.
Brawn has been replaced time and again and employment recovered.
This time might be different, AI being the difference.
A time will come where employment won’t recover or only at much reduced salaries.
It’s all deflationary, especially if demand is reduced with the jobs.
Efficiency savings can be illusory if cost savings have to be passed on to maintain demand levels in a reduced salary and employment market.
Exactly! It’s all deflationary but the central banks abhor deflation. The cost of living should be dropping exponentially, like the price of 4K UHD TVs, but the Keynesian klowns insist that prices must rise.
Plenty of office book keeping positions have been destroyed. I think brains have been replaced at a faster and faster rate. I think the work week eventually changes to compensate just as it had in the past to 3 and 4 day work week. The new jobs already exist, they are just way different than one would expect. I also see a big change to universitiy education coming. We are hiring more apprentices at minimum wage while they go to college because a undergraduate degree doesn’t mean anything to us when we interview anymore.
As more human jobs are replaced by automation, humans need to curb their reproductive activity to match the ability to support a family.
We have watched “invincible” high frequency trading algorithms get hacked. We have heard of people hacking into the websites of the FBI, CIA, etc. We have seen people hack into ATMs, traffic cameras, impregnable cell phone systems, MSwindows network controls, yahoo / google emails, the DNC email system (the one in DC – not Hillary’s which had no security). Every week, there is another data breach at a major retail chain. The military has lost track of tens of thousands of military personnel records — on lost USB drives when the data supposedly can’t be copied at all much less to a USB. The CIA allegedly has turned every Samsung TV into a spy tool (and if they haven’t, it won’t take long for someone else). And not to be left out, someone (other than Edward Snowden) apparently leaked a new batch of the CIA’s dirty laundry to Wikileaks.
Now is Mish seriously going to tell us that these truck convoys can’t be hacked and turned into a trial lawyer’s dream case? And that’s just for money.
Every crack pot “activist” group with a few bucks from George Soros is going to hack these trucks and drive them into gas stations, planned parenthood, pet stores, and everything that offends the snowflake generation (which is everything). And it won’t be long before the snowflakes turn on their masters — broken by student debts to finance Krugman’s retirement kitty — and hack one of these 100ton wrecking machines through the front gates of the universities that enslaved them.
Mish is so fixated on the wiz bang technology that he forgets the stuff has to fit in to society. Truck companies probably want to reduce / eliminate the cost of drivers, but it is naive to think that is the only cost here.
Even if one is too old/clueless to realize the hacking issues, there is also an economic problem. Unemployed truckers aren’t going to pay for all the cheap imported crap these trucks are intended to carry.
I don’t know why Mish is so fixated on driverless technology, but wish he would think about the entire implementation instead of whether software can drive as well as a DMV official. There are many cultural issues that Mish has completely ignored.
Yeah, yeah. The Titanic was unsinkable, and these systems will be unhackable.
I’ll bet the legal profession is busy arranging nationwide seminars regarding the newly formed division of class action lawsuits, death by driverless truck.
Watch for the non stop adds on TV at the noon hour and all morning long.
Good call. Insurance costs alone will make self driving trucks useless for all but the most limited routes.
Exactly wrong. Insurance for human-driven vehicles will make that activity cost-prohibitive.
Self-driving technology will eliminate almost all road fatalities.
Three big rigs barrelled up and down the 110 Freeway on Wednesday…”
But wait, that’s not all. Speaking of California roads, today’s Sunday L.A. Times has a story titled “Robot cars set to hit roads in 2018.”
Friday, the DMV released a new proposal for regulations for the testing and operation of driverless cars. The story notes that the new rules are friendlier to testing but some restrictions still apply.
To the person whose lengthy and mostly nonsensical comment that I deleted, I am not backing away from anything. My essential idea mentioned many, many times over the years was long-haul trucking would vanish first, and for some time there may be a driver needed to take the truck to its final location.
Should it play out this way, over 90% of the miles will be totally driverless.
I suspect I am overly pessimistic given self-driving taxis will be available by 2022 (but perhaps not widespread).
I’ve spent the last three months in Mountain View, California, which has given me the opportunity to frequently see Google/Waymo’s self-driving cars in operation first-hand. What I’ve seen is an impressive achievement that is nowhere close to being ready for widespread adoption.
First, the entire fleet of self-driving cars reliably disappears at the first sign of poor weather. And by “poor weather”, I mean rain. A moderate amount of rainfall that no human would think twice about driving in shuts down the self-drivers. According to the people I’ve talked to, the high level of moisture in the air tends to scatter and degrade the LIDAR sensors to the point where it is no longer safe to operate the cars.
The cars are still very poor at handling unexpected traffic situations. On Friday, I saw one of the “bubble” cars encounter a large auto transporter parked in the bike lane at the side of the (2 lane each direction) road near the Toyota dealership on Middlefield Road. The auto transporter was almost entirely clear of the travel lane, hanging maybe a foot into the lane. The self-driver spotted this, came to a full halt and refused to proceed. There was plenty of space to pass the transporter by moving to the left side of the right-hand travel lane, and the left-hand lane was also completely empty. I spotted this ahead of me, switched to the left-hand lane and blew on by.
The old joke in the software industry is that the first 90% of the project takes 90% of the time, and the last 10% takes the other 90% of the time. I’d say that today’s self-driving vehicle technology is probably about 70% of where it needs to be to become viable. Some of the problems could be overcome by equipping the roads with markers and traffic signals that are friendlier to self-drivers (at a huge cost).
The current technology also relies heavily on high-resolution and highly accurate maps of the areas that the cars operate in. Mountain View California is probably the most accurately mapped city in the world, and bringing the rest of the world up to that level of accuracy will be a massive undertaking. Maintaining that accuracy over the long haul is a formidable challenge.
Self-driving vehicles that can reliably navigate urban environments under all possible weather and road conditions are decades away. I think we will see this technology start being employed in heavy trucks as driver-assist safety systems in the next five years, but it isn’t going to completely replace drivers any time soon.
Nice phrase “under all possible weather and road conditions” are decades away.
Even with that stringent out, you are likely far wrong.
But we do not need “all possible” before millions of jobs disappear. We need about 5 years or less at the current pace of progress and competition.
I’m curious what level of performance you think will be adequate to deploy this technology? I don’t think it’ll be necessary for the self-drivers to be able to operate in the middle of a snowstorm, but I do think they’ll have to figure out how to operate in the rain.
What we have today as “self-driving vehicles” are what I’d describe as hand-built engineering prototypes. IF the technology that exists TODAY is good enough, it will still take over five years to productize that technology into a production-ready sensor & electronics package that can be manufactured in large quantities and be mounted on existing vehicle designs, or to modify the upcoming vehicle designs to accommodate the self-driving package. To have a measurable impact on employment, there will need to be 10’s of thousands or 100’s of thousands self-driving vehicles in existence and ready to roll.
It’s not impossible that somebody will develop a prototype self-driving vehicle in the next five years that WILL be “good enough” to do the job. I personally think it will take more like 10 years. Once we hit the “it works” milestone, it will take another 5 to 10 years to turn that prototype into a large number of ready-to-roll vehicles.
The reason for my 10-year estimate for the technology development is that operating a vehicle in human-operated traffic is a complex matter of non-verbal and visual communications between drivers. A human operator may note that the car in the left lane is a sports car that is being aggressively driven, and it would be smarter to let it go by before moving into the left lane. The self-driver is more likely to decide that it could safely enter the left lane and get rear-ended because the other driver wasn’t paying attention.
If placed in the middle of the chaotic and aggressive driving styles in places like Boston or New York City, I predict that the current crop of self-driving cars are not going to do well. The day I see a prototype self-driving car drive unscathed through Boston during rush-hour without human intervention, I will start the 5 to 10 year clock ticking on those jobs going “poof”.
It would happen quicker if there was a way to instantly convert ALL the cars to self-driving overnight.