LONG BEACH, Calif. — With testing of its Mark 1 prototype complete, Parallel Systems has turned to building and testing the next iteration of its autonomous rail initiative.
And at the Intermodal Association of North America (IANA) annual meeting here, Parallel CEO Matt Soule said the startup’s goal is not just to make existing rail capacity autonomous and battery-powered. It is to displace trucks on some short-haul to medium-haul lanes, with a particular focus on staying ahead of gains being made in autonomous trucking.
Parallel raised almost $50 million earlier this year in a series A funding round. Soule said it would keep the company operating until 2024. The key principals at Parallel came out of SpaceX.
“How do we leverage the rail network that is out there, and where can we take advantage of the technology that has been developed by electric vehicles and the self-driving industry,” Soule said in an interview on the exhibition floor of IANA. “We are looking at driving more business to rail.”
Mark 1 was a prototype that used two robotlike vehicles equipped with train wheels upon which sat a container. The pair of carriers were at each end of the container.
A 3-mile stretch of track operated by Parallel in Fillmore, California, with no other traffic and no grade crossings, has been used as the test ground for Mark 1. Soule said testing of Mark 1 was completed this past spring.
The next step, Mark 2, looks more like a transport system designed for a container. It is one piece resembling a chassis, and a container sits on top of that. At the IANA meeting, the Parallel booth played a video that showed the Mark 2 carrier being built and moving its first few autonomous feet on an internal track.
“We are in the earlier stages of integrating and testing the second-generation vehicle,” Soule said. “We are working with the [rail] industry, going through a lengthy test program. We are looking forward to operating on a real network sometime in the future.” He did not disclose specific time frames.
Both the Mark 1 and Mark 2 transport systems use autonomous EV technology for propulsion. One difference: Because weight is not as much of an issue with rail as it is with trucks, the battery is a lithium-iron phosphate battery, which is heavier than a lithium-ion battery but cheaper as well.
The target customer for the Parallel product will be railroads, Soule said. That doesn’t mean they need to be big boys like BNSF or Union Pacific. They can also be short-line railroads, looking to reach customers that are not now easily served by rail.
Drayage is the underlying theme for the IANA meeting’s almost 2,000 attendees, but Soule stressed that drayage is not the sole application for Parallel’s technology.
“Rail generally serves one corner of the market,” he said, describing that sector as “high volume going long distances.” That leaves lower volume and shorter distances to be dominated by trucking.
“With drayage, costs dominate, so why don’t I just put it on a truck?” Soule said of the prevailing economics. He believes the Parallel system could have a port increase the number of drayage truck “turns” in a day from as few as three to five or six.
But Parallel’s vision is that with the lower costs of an autonomous rail carrier, with no locomotive and with the efficiency advantage of rail versus trucks, “this is a tool to allow those railroads to go after those smaller markets,” Soule said. The scale required to make conventional rail efficient is not required in the Parallel vision. And with that lower cost structure, Soule said, smaller terminals can be opened that could be served by an autonomous rail system.
The fabrication of the Mark 2 was conducted by an outside company, Soule said. The 50-person staff at Parallel at this point is driven by creating the autonomous vehicle software that ultimately will direct the next generation of Paralell’s product. Parallel’s software will also have the capability to ask for “authority” from the system that manages the various railroads the company eventually hopes to put its products on.
It doesn’t have a name yet, but Soule said Mark 2 is not expected to be the commercial offering that would be brought to market. Parallel will learn from Mark 2 and the next product will be the commercial offering. There is no firm timeline on when that will be. (For example, the body of Mark 2 was mostly constructed by steel welding. Soule said he expected the final commercial product to have a cast product for a chassis.)
The sensor package that would be part of the Parallel offering has been tested on existing tracks, Soule said. It was installed on a locomotive for an unidentified railroad, but did not have any role in guiding the locomotive. But in doing so, Soule said, Parallel was able to test the ability of the sensor system to do exactly what its name suggests: sense the presence of cars at crossings, people walking nearby and changes in direction of the tracks.
But Soule came back to one of the arguments Parallel is making about autonomous rail versus autonomous trucking. Autonomous rail will be on tracks only with the permission of the owner — possibly having been contracted by the railroad to move freight — so the vehicle’s sensor system can be built with knowledge of the track. An autonomous trucking system must make do with whatever is on the road.
“Accounting for their central knowledge of how this railroad is operating is leverage for us,” Soule said.
A train of vehicles like the Mark 2 could be 50 cars long, or about a half-mile in length, Soule said. But he suggested that a 10-vehicle train might be a sweet spot for a Parallel train.