Automated technology in trucking and logistics in the U.S. may be steps ahead of the infrastructure and regulatory environment needed to support and advance what could be the future of freighting and logistics.
Even as venture capital pours into driverless trucking startups, some workers have stepped up to protest technology that could reduce the rank and file amid efforts by the U.S. Department of Transportation (DOT) to outline the road ahead for a safe approach to automation in mobility.
To date, some of the most promising automated trucking startups have attracted hundreds of millions in backing from venture capitalists banking on driverless technology as the future of an industry in which long-haul drivers face grueling hours and, importantly, a critical shortage of peers. According to the American Trucking Associations, the industry needs 50,000 more drivers though that figure swells to more than a quarter of a million when looking at numbers from FTR Transportation Intelligence.
The domestic trucking industry employed 1.75 million heavy and tractor-trailer truck drivers, according to data from the U.S. Bureau of Labor Statistics (BLS) Occupational Employment Statistics, in addition to 877,670 drivers operating light or delivery services trucks, and 427,000 people classified as drivers/sales workers. Sixty-one percent of all freight that moved throughout the U.S. in 2016 was transported by truck, per BLS data.
Under pressure to keep costs low, trucking operators could reduce expenses by switching to automated technology. A PwC study indicates that automation could result in annual savings of 28 percent for the operating costs of the typical long-haul truck.
Beyond the cost concerns, fledgling companies innovating in automated truck technology believe their advancements could make truck driving safer, not just for employees in the industry but for everyone else on the roads, too. This is a significant factor in why startups including TuSimple, IKE, Apex.Ai, Boxbot, Kodiak Robotics and Embark together have raised a total of $340.2 million—and TuSimple is the rare unicorn in the group, with a valuation north of $1 billion.
Though all of these companies are working toward the end goal of automated or driverless truck operation via systems largely powered by artificial intelligence (AI), machine learning (ML), computer vision and a wealth of sensors, their individual approaches vary.
Embark equips semi trucks with its technology to enable automated highway driving. CEO Alex Rodrigues, who studied mechatronics engineering at the University of Waterloo, claims Embark currently operates the world’s longest automated freight route. Together with partners Frigidaire and Ryder, Embark trucks travel autonomously over the 650 miles from Texas to California. Human operators move the truck beds from the warehouse to the highway’s edge, where an Embark rig attaches and sets off on the journey, with the reverse process occurring once the trip is completed. An Embark truck bearing the Amazon logo has also been spotted on the road.
TuSimple seems to go a step further, claiming its high-tech trucks can safely navigate both open stretches of highway and congested municipal streets. The company currently has 12 contracted customers and plans to increase its fleet of 11 fully autonomous trucks to 50 by June to meet demand from Fortune 100 and international firms.
In a statement earlier this year, TuSimple founder and CTO Dr. Xiaodi Hou said, “Exactly one year after debuting our prototype system at CES 2018, we’re now running up to five commercial trips a day in Arizona, expanding our fleet and moving quickly toward our goal of creating the first commercial self-driving truck.”
Trucks in TuSimple’s fleet leverage a “camera-centric perception solution” that offers visibility to 1,000 meters, which the company claims is superior to other offerings in use by competitors. In February, the company announced a $95 million Series D round that brought it past the $1 billion mark and said it will use the investment to further develop its product and ramp up commercial readiness. In addition to the fully autonomous deliveries it makes daily in Arizona, TuSimple said it will have similar operations in Texas shortly for large shippers and fleets.
Colin Xie, vice general manager, investment department for Sina Corp., which led the latest funding round, said, “We are focused on finding the global leaders in artificial intelligence and TuSimple is ahead of the pack.”
Kodiak Robotics raised $40 million in August in a Series A round led by Battery Ventures with participation from CRV, Lightspeed Venture Partners and Tusk Ventures. The autonomous-vehicle startup’s leaders boast considerable experience in the nascent driving automation space. Founder and CEO Don Burnette, who had worked on the Google team developing self-driving automobiles, previously co-founded Otto, the self-driving truck startup that was acquired by Uber and subsequently shuttered last summer.
“We believe self-driving trucks will likely be the first autonomous vehicles to support a viable business model,” Burnette said in a statement regarding the venture investment.
Itzik Parnafes, general partner at Battery Ventures, added, “Autonomous driving is likely one of the most major technology shifts of the last 100 years. We have been researching this trend for several years and feel confident we are backing an experienced, savvy team that can capitalize on the opportunity in a unique way.”
Startup Apex.AI creates automated driving technology for mobility applications. Apex.Ai’s offerings serve everything from passenger automobiles to automated trucking and logistics. Boxbot, whose team includes experienced engineers and veterans from Uber, Tesla and Amazon, focuses on where automation can improve operations in last-mile delivery.
Brian Wilcove, a partner at Artiman Ventures, which led a $7.5 million seed funding round last June, said, “Over the next few years, self-driving vehicles will transform the last-mile, making it cheaper to make deliveries and easier to receive them.”
Given all of the innovation and investment funding flowing into the driverless trucking technology space, the DOT is working to ensure regulatory agencies and all ancillary stakeholders take the necessary steps to ensure the U.S. is prepared to accommodate full-scale automated trucking on its roadways. The DOT issued the Automated Vehicles 3.0: Preparing for the Future of Transportation report last fall to outline the challenges that federal bodies, states and local municipalities should address to pave the way for automated truck driving.
The transportation body “is taking active steps to prepare for the future by engaging with new technologies to ensure safety without hampering innovation,” the report said. It’s also adopting a safety-first and technology-neutral approach to automated transportation.
Based on recent protests, some workers aren’t looking forward to the arrival of automated truck technology, which they see as a threat to their livelihoods. Local news agencies reported that hundreds of Port of Los Angeles union dockworkers banded together to voice their opposition to the use of automated trucks at a terminal they service, fearing that the technology could replace or eliminate their jobs. Dutch logistics firm APM applied for a permit to operate driverless cargo vehicles in its terminal, a motion that was tabled by the L.A. Board of Harbor Commissioners.
In a March 15 letter from APM to the commissioners and the Harbor Department director of planning and strategy, APM noted that “automation is expressly provided for under the Collective Bargaining Agreements between the ILWU and the Pacific Maritime Association” and its zero- or reduced-emission technology is consistent with initiatives like the 2017 Port Clean Air Action Plan Update and the State of California Sustainable Freight Action Plan.
Despite the uproar, recent developments indicate California’s open-mindedness toward driverless technology. On April 12, the state’s Department of Motor Vehicles proposed regulations that would allow companies to test and deploy autonomous delivery trucks weighing less than 10,001 lbs. on public California roadways.
Automation promises enhanced safety and efficiency in trucking but more work needs to be done to ensure that the highly complex AI and ML systems that power driverless trucks are secure from adversarial attacks. Dawn Song, a UC Berkeley professor and expert in AI and ML security, spoke at MIT Technology Review’s EmTech Digital event about potential risks that could threaten self-driving systems, as the publication first reported.
Just as AI and ML algorithms and systems are trained of vast troves of data, bad actors could input information into these systems to reveal the information they’re built on—thus uncovering their exploitable weaknesses. Song spoke of a test in which stickers applied to a stop sign fooled a vehicle’s computer vision system into thinking the signage actually outlined a 45-mile-per-hour speed limit, which could have disastrous consequences for driverless vehicles. Such “adversarial machine learning” needs to be addressed before fully automated fleets take over the roadways.