Autonomous trucks hold more immediate promise for deployment than robotaxis, industry experts say. Compared to complex city roads—where a myriad of challenges abound, from navigating heavy congestion to watching out for jaywalkers—highways and lonely ports are relatively easy to navigate.

While a number of companies developing autonomous trucks are focusing on full autonomy, it’s not strictly necessary. More immediate applications lie in the realm of semi-autonomous rigs.

China-backed Tusimple is one of the pioneers of highly autonomous trucks, which can drive themselves under certain road conditions. Founded in 2015, the San Diego- and Beijing-based startup has secured $178 million in funding. It is now worth a whopping $1.1 billion, making it the world’s first autonomous trucking unicorn.

Late last year, Chen Mo, the CEO of Tusimple, said the company was working at “almost the same speed” as Waymo in terms of commercialization.

The claim may sound overconfident, but Chen was later vindicated. This year Tusimple won a contract from the United States Postal Service (USPS) to carry letters and packages between Phoenix, Arizona and Dallas, Texas—a 1,600-kilometer trip. It may only have been a two-week pilot, but it marked the first time that USPS contracted with an AV company for long-haul services, and it gave Tusimple, a company with Chinese roots, a chance to validate its system with a US government agency.

Drive I/O

Drive I/O is TechNode’s monthly newsletter on the cutting edge of mobility: EVs, AVs, and the companies trying to build them. Available to TechNode Squared subscribers.

Like Waymo, Tusimple aims to reach full autonomy and plans to run a fleet of 1,500 cargo-hauling trucks by the end of 2021. Hou Xiaodi, founder and CTO of Tusimple, said earlier this year that it is now on track to take “drivers out” of the trips starting next year.

The company claims that its proprietary deep-learning algorithms could enable vehicle to “see” the road a kilometer ahead, which sets it apart from other AV companies, including Waymo. Tusimple equips each vehicle with nine cameras, two lidars, one forward radar, and an onboard computing platform. All this technology costs around $200,000 per vehicle.

On top of these costs, the company employs around 400 staff members focused on engineering and marketing. Each truck now earns several thousand dollars on average each week, Chen said, though this does little to help profitability.

Aiming to reduce costs and kickstart commercial use, the company is looking at C-V2X (cellular vehicle-to-everything) for communications, a networking solution for vehicle connectivity that the Chinese government has put its clout behind.

The technology allows sensors and software to be deployed on road-mounted devices, which could reduce the costs of autonomous trucks. Data from the road could be sent to robo-rigs and combined with the trucks’ own data in real time. The system will allow detection of traffic scenarios that onboard sensors are currently relied on to complete.

China plans to use wireless communication solutions and sensors for connectivity on 90% of highways in the country by 2020. However, C-V2X is still a nascent technology with a multitude of issues that need addressing, including how to transmit data from road signs to driverless vehicles, as well as real-time onboard computing.

Revamping public road infrastructure also requires substantial amounts of money, meaning that only major Chinese cities are currently able to make the investment. Also, the technology used in one city may not be compatible with vehicles from other regions.

This would restrict the technology from being used on long-distance trips, said Wu Nan, the vice president of Tusimple, during a panel discussion at this year’s Mobile World Congress in Shanghai. Thus far, the company sees little hope of any practical applications of C-V2X in the near-term.

Finally, concerns from lawmakers and the public are the biggest obstacle to realizing the vision of completely autonomous trucks on the road. Currently, the US laws regulating highly autonomous driving require that a driver be behind the wheel at all times, ready to take over operations if needed. Meanwhile, public road testing for trucks hasn’t even started in China.

While driver shortages in the US are likely to double to 160,000 in the next ten years, autonomous trucks could replace up to 300,000 long-distance truck drivers in the US over the course of the next two decades. Society will have to make some hard choices between protecting current livelihoods and maximizing incentives from new industries.

Level 3: Back on the table

As the industry struggles with technological issues and money gets tighter, support in China is growing for another approach: conditional autonomy.

As the name implies, a truck with conditional autonomy—also known as a Level 3 vehicle—can drive itself under certain traffic or environmental conditions, but still requires a driver to take the wheel if necessary. Admittedly, the concept sounds less thrilling when compared to Level 4 and Level 5 technology, in which the vehicle’s performance is equal to that of a human driver.

Nor is conditional autonomy without its concerns. Auto giants such as Mercedes and Volvo question the safety of handing over control from vehicle to human driver, as well as the effects of the continuous transitions on the driving experience. Nevertheless, because full autonomy is still a long way off, companies such as the Chinese autonomous truck startup Inceptio believe that Level 3 capability is a mandatory stage of the process to transform the logistics industry.

China is a country of diverse climates, landscapes, and traffic conditions. These conditions require truck drivers to be highly skilled, with wide-ranging knowledge of routes, vehicles, and even the cargo that they haul. Julian Ma, Inceptio CEO and a former Tencent vice president, has said that the role of truck driver is a manual job, and there is no way to increase profits under the current non-standardized business model.

Founded last year by Tencent-backed fleet management company G7, Asia’s largest warehouse operator GLP, and Nio-backed venture capital firm Nio Capital, Inceptio unveiled the first generation of its autonomous driving solution in Shanghai at this year’s CES Asia. The company claims that its “full-stack solution” could enable a truck to deal with multiple complex tasks including automatic braking, lane changing, and even U-turns in different situations.

Once they begin to operate on a large scale, semi-autonomous trucks could standardize the industry, allowing an inexperienced yet licensed driver to undertake any trip, Ma says. He added that logistics costs could be reduced by at least 10% and that drivers will not have to work long hours in poor conditions, as long-haul trips could be divided into several relay-like segments. The company aims to bring its customized robo-rigs to market at scale by the end of 2021, running a delivery network with a fleet of at least 50,000 vehicles nationwide.

Although the Level 3 goals are more realistic, the developers working on Level 3 technology still face a variety of challenges that would never be encountered by robotaxi companies.
One such issue is related to automobile controls.. Since cargo delivery involves the circulation of shipping containers, sensors must be installed on the truck tractors themselves and combined with algorithms of vehicle modeling and controller design that help vehicles cruise along the roads.

Also, the weight and position of the cargo inside the container has an effect on the motion of the truck, making it harder to control during the trip.

Engineers need to test specific problems and corresponding solutions, taking into consideration a multitude of other factors, including tire pressure and the roughness of the road. Currently, Tusimple is still unable to transport liquid goods because sloshing within containers shifts weight, said Chen.

Despite these challenges, Inceptio is still convinced these issues can be solved with time and effort. The company is currently testing autonomous trucks in Shanghai, the northern Chinese city of Baoding, and Changsha, the capital of Hunan province. Changsha is expected to open 200 kilometers of highways and urban roads for autonomous tests by the end of the year. Inceptio, which holds one of the two permits awarded to autonomous trucking companies, believes that the ability to test and troubleshoot in Changsha will give it an edge in the race towards the mass use of autonomous trucks in China.

Jill Shen

Jill Shen is Shanghai-based technology reporter. She covers Chinese mobility, autonomous vehicles, and electric cars. Connect with her via e-mail: jill.shen@technode.com or Twitter: @yushan_shen

Chris Udemans

Christopher Udemans is TechNode's former Shanghai-based data and graphics reporter. He covered Chinese artificial intelligence, mobility, cleantech, and cybersecurity.