'Hundreds' of Gatik Robot Delivery Trucks Headed For US Roads

An anonymous reader quotes a report from Forbes: Gatik, a Silicon Valley startup developing self-driving delivery trucks, says its commercial operations are about to scale up dramatically, from fewer than a dozen driverless units running in multiple U.S. states now to hundreds of box trucks by the end of the year. CEO Gautam Narang said it's also booked contracts with retailers worth at least $600 million for its automated fleet. "We have 10 fully driverless, revenue-generating trucks on public roads. Very soon, in the coming weeks, we expect that increase to 60 trucks," he told Forbes. "We expect to end the year with hundreds of driverless trucks -- revenue-generating -- deployed across multiple markets in the U.S."

Though the Mountain View, California-based company hasn't raised as much funding as rivals, including Aurora, Kodiak and Canada's Waabi, Gatik said it's actually scaling up faster than any other robot truck developer. Unlike those companies, it focuses on smaller freight delivery vehicles, rather than full-size semis, supplied by truckmaker Isuzu that operate mainly between warehouses and supermarkets and other large stores. The company's focus has been on so-called middle-mile trucking, which, like long-haul routes, has a severe shortage of human drivers, according to Narang. Currently, its trucks are on the road in Texas, Arkansas, Arizona, Nebraska and Ontario, Canada.

The company has been generating revenue since shortly after its founding in 2017, hauling loads for customers like Walmart in trucks with human safety drivers at the wheel. Beginning late last year, it began shifting to fully driverless units and is getting more trucks from Isuzu built specifically to incorporate its tech, Narang said. "The hardware that we are using, this is our latest generation, has been designed to enable driver-out across thousands of trucks."

Why focus on the last mile?

[n-baxley]

The last mile of a trip seems like the worst place to employ driverless vehicles. You have the most variety of roadways, delivery docks, other staff, so many ways to go bad. Why not focus on interstate driving and take on the long boring stretches that require trucks large enough to sleep in?

Re: Why focus on the last mile?

[mcgrawcm82]

Totally agreeâ" who the hell is going to unload all the merchandise? That used to be the truckerâ(TM)s job, am I right? Now stores are going to have to employ someone to unload the trucks because theyâ(TM)re driverless.

Re: Why focus on the last mile?

[sziring]

Stores frequently unload the trucks themselves anyway. The driver usually just drives.

Re:

[drinkypoo]

Stores frequently unload the trucks themselves anyway. The driver usually just drives.

Sometimes even while they are still unloading.

You could make trailers that unload themselves. You'd need to load them more intelligently than is usually done, but it is possible.

Re:

[noshellswill]

During truck deliveries by free-lance owner-operators the unloading is often performed by an unofficial, but well organized "grey market". Workers "appear" at the unloading site ( unemployed/illegals/felons/2nd-jobbers ) and do the merch unloading; they are paid cash by the owner-driver;  very "leaky" transfer, but no tax no bonding no security-check no union.

Re: Why focus on the last mile?

[sziring]

Because it has the least amount of competition currently. They stand to gain a lot. Next step is to use a drone or bot to drop off packages at the front door. It's a brilliant play as Amazon, UPS, FedEx etc. all want this ability.

Re:

[j-beda]

The last mile of a trip seems like the worst place to employ driverless vehicles. You have the most variety of roadways, delivery docks, other staff, so many ways to go bad. Why not focus on interstate driving and take on the long boring stretches that require trucks large enough to sleep in?

It does seem like a business model where human drivers take it from the loading dock to the highway to let it drive itself to the next city where it is met by a local driver to take to to the next loading dock would make a lot of sense. The local drivers could likely keep reasonable working hours (along with the people running the loading/unloading facilities) while the trucks run 24/7 and wait in a big parking lot near the highway.

Then again, spending resources setting up this system and fine tuning it for

Re:Why focus on the last mile?

[DarkOx]

because that isn't where the costs are.

Shipping lane management is already highly data driven and pretty well optimizes. Sure you might save some money replacing long haul drivers, and independent owners, with drone trucks, it is not insignificant, but it also isn't probably the big part of the pie.

I don't know about Walmat, given the volume of stuff they have to move to keep stores stocked, and sell in store; but I would be shocked if Amazon spends more money moving merch between distribution hubs than on last mile. Think about how many people are doing hand-pick/sort, loading trucks, and driving around all day to move relatively small volumes of stuff.

Look at it this way, millions of dollars in merch fit on 53' trailer, and the driver costs you maybe $400 a day. Running the truck costs much more than that but you still will need fuel, insurance, maintenance etc even without the driver. You last mile driver earns a lot less he maybe costs $250 a day but he is moving maybe what 10k worth of merch around on a good day?

From a technical standpoint sure the over the horizon guys are probably easier targets for replacement, and its a huge industry but for the biggest investors in this stuff last mile is their first target because it is bigger cost center for them at least.

Re:

[SumDog]

I'd guess about half of wal-mart's freight uses spot markets; literally do to day pricing. Everyday, logistics people at major retailers and freight carriers are in a constant bidding war to secure trailers and freight. Some stores, like Dollar General, use up to 50% dedicated (I want 300 tractors and drivers from company x, every day). You want to keep dedicated under your max by a lot. There's nothing worse than having drivers sitting there, doing nothing, and still making rates. But everything after that

Re:

[havana9]

Trains, ships and planes are well optimized for this, the cost of having drivers is a small fraction of the overall costs. Once the stuff it's loaded in a container there's already an optimized system in place.

Re: Why focus on the last mile?

[larwe]

Did you read the article? This explicitly is not last mile (distribution to individual customers) it is middle mile (warehouse/distribution center to store). These trucks arenâ(TM)t carrying 1000 individual Amazon packages that need to be delivered to 1000 individual addresses, they are carrying half a dozen pallets of shrink wrapped merchandise, all of which is going from the same point A to the same point B. The truck pulls up, door opens, someone with a pallet jack unloads it in 2 minutes.

Re:

[SumDog]

Because that requires legal sign-off from multiple states, probably at least 5~7 for any type of OTR lanes. No state is going to approve that at such an early stage, and railroads are just so much better at that anyway. A lot of UPS/Fedex trailers still get loaded directly onto flatbeds.

Re:

[evil_aaronm]

I wish rail were more accessible. I deliver typically two pallets of goods from Podunk to 3/4 out on Long Island - it's about 450 miles one way - and it would be grand if I could stick it on a train and have it delivered. But there's no real terminal for trains either here or there. As it is, we use LTL when we can, which I suppose is just as good, but having driven through the NYC metro area hundreds of times to make this delivery, trains would help minimize commuter traffic.

Re:

[SoftwareArtist]

That's why they're focusing on middle mile, not last mile. Last mile is hard. Your truck needs to go anywhere, and you don't know if there will be someone there to unload it at the end. Middle mile between warehouses and stores is a lot easier. It only needs to drive fixed routes that you assign in advance, and you know there will be someone to unload it at the end.

Why not focus on long-haul highway driving? Because other companies are focusing on that. This is a different market that requires a diffe

Re:

[CommunityMember]

Amusingly, one of the firms located on the other side of Evelyn/Central Ave from Gatik's Mountain View office is doing that (long haul), i.e. Aurora. There is a lot of middle mile (warehouse/depot to loading dock) deliveries in the US (go around to the back of your larger grocery store and look at the trucks; some may be semi-sized, but many may be the classic box truck). Middle mile is a bit harder than driving across the country on I-80, but both need to end up pulling into the loading docks at both end

Loading and unloading

[memory_register]

More than half of last mile delivery is unloading. Do they think a store owner is going to unload all their bulk groceries, or grandma is going to unload that new washing machine?

Re:

[SumDog]

Have you worked at loading docks? Most drivers don't unload trucks, unless they're full time employed by the receiving freight company (FedEx, UPS, etc.)

Re:

[lucifuge31337]

Drivers don't unload trucks except for in very specific situations like bread and snack truck routes where they are maintaining a retail display.

Re:

[evil_aaronm]

It depends. I receive goods that are palletized, and the driver grabs the pallets off the truck and dumps them on our dock. It might take him all of a minute per pallet, depending on how well the rest of his load is situated, but he does it.

Re:

[lucifuge31337]

"Grabs" the pallets with what? A pallet jack that he keeps in his truck? That's very much not what is being talked about in what I responded to. You are not getting drivers pallet jacking skids of grocery items out of a trailer. In fact, it's much more likely they will drop a trailer on one dock and pick up the empty on the other dock in these types of situations.

Re:

[evil_aaronm]

Well, in fact, 40 years ago when I worked as a stock boy in the local grocery store, we *did* get pallets of grocery items out of a trailer. The driver and shift manager took turns unloading the pallets from the trucks. I know this, because I either watched, or took our jack and moved loaded pallets out of the way, or took product off the pallets to stock on the shelves. Got an interesting story about another stock boy who didn't know how to work a jack, but that's for another time.

I don't know if that

Re:

[dvice]

They can use stretch the robot for loading and unloading:
https://bostondynamics.com/pro... [bostondynamics.com]

And In related news

[Anonymous Coward]

Highway robberies see a huge resurgence since the introduction of Robot Delivery Trucks.

Finally.

[gurps_npc]

I have no idea how the idiots decided to focus on consumer cars that go everywhere.

Trucks and busses have set routes. This example - the middle mile between warehouses and individual stores - is the prime case for robotic delivery. Both locations have employees working for the same company to handle the loading and unloading. The routes do not change. Speed is needed but not absolutely necessary. Most of the route should be on highways.

This is where the first commercial unmanned vehicles should be used (rather than 'testing' versions that need permission we currently see).

Re:

[drinkypoo]

Trucks and busses have set routes.

They have to be able to handle detours.

Most of the route should be on highways.

Some of these trucks will spend most of their time on surface streets, and have to navigate more obstacles there.

This is why long haul trucking was first and most applicable. It's least likely to involve substantial surface street traffic because a lot of it is warehouse to warehouse.

Re:

[Gravis Zero]

I have no idea how the idiots decided to focus on consumer cars that go everywhere.

That's actually a very simple question: there is far less risk involved. The bigger problem with driving is pedestrians. It may seem easy to not hit people but that's not the case. Having an autonomous car hit someone at low speed is a serious problem for a company. Have an loaded truck/bus hit someone at low speed is typically fatal and a surefire way to get your vehicles banned.

In sum, the problem is not navigating the route, it's doing it without killing people and trucks/busses are far more likely to ki

Re:

[CommunityMember]

I have no idea how the idiots decided to focus on consumer cars that go everywhere.

For some, the long haul and middle mile market is just a subset of the much harder problem (taxis). If they can handle consumer travel (and in some locations, Waymo seems to have mostly solved that problem), they can enter the long haul and middle mile market whenever they see it as a profitable market to enter. Aurora (long haul) and Gatik (middle mile) have an advantage today in those markets because the have focused on just those markets.