Torc is an autonomous long-haul trucking company with a focus on safety and integrating with the existing freight network.
You arrived at Torc from Daimler only 6 months ago; can you tell us what Torc does and what is its relationship to Daimler?
Torc is 17 years old and received a majority investment from Daimler Truck in 2019, operating as an independent subsidiary and the cornerstone of Daimler Truck’s autonomous strategy. Over the years, we have seen the value and market need for bringing a safe and transformative autonomous trucking product to market. With this in mind and working towards this goal, Torc has grown from a small Virginia-based company to four locations globally across four time zones, most recently acquiring Algolux, a computer vision and machine learning company based in Canada. Torc executes each step with its mission in mind: creating safe technology for on-the-road driving and testing.
You claim your autonomous driving system “sees-thinks-acts”; what does this mean technically?
Teaching an autonomous truck to perceive the environment is a really difficult thing, as all sensors have deficiencies. Our system consists of roughly 30 sensors that not only allow the vehicle to see, but also to localize the vehicle in real time. With that said, the majority of the sensors are visually focused, using cameras, radars, laser scanners, and more in order to maximize the system’s efficiency.
Through these, the truck can orient itself in terms of location and context (where is the truck and what lies around it, for example, a parking lot with an object behind, etc.). In terms of ‘thinking’, the system has the capabilities to predict what the objects around it will or might do in the immediate future, and thus decide how and where it should move. This is decided by a mixture of artificial intelligence (AI), machine learning, and traditional robotics software. The execution or ‘act’ is basically recreating the movements that the driver would typically do to operate the vehicle (such as pumping the brakes or turning the steering wheel) with wired signals, taking into account a wide set of variables that may occur.
In unforeseen circumstances where the secondary system fails or the electric support in the steering wheel fails, the driver of current trucks has to apply force to steer. In a truck integrated with autonomous driving systems, all of this is instead done by actuators. This ‘see-think-act’ approach might sound simple at first, but in reality, it is a super complex and interesting challenge that our engineers are up to solving.
When do you think autonomous trucks will be available at scale?
Of course, I might be rather biased, but I do think that autonomous trucks will be available in the U.S. at scale by the end of this decade. It will probably start with slow volumes by the middle of the decade, and then grow step by step. This is the timeline for the U.S., which by virtue of several factors is set to be the first country to adopt autonomous trucks. The economic strength, the incredible talent, the innovative spirit of its people, and the amount of capital going into development and technology here really make the country a forerunner.
Additionally, with the current labor shortage of drivers in the U.S. touching 80,000, autonomous trucking would allow for less time away from home for drivers and a more flexible career path for those considering entering the industry.
In comparison, more complications may occur in Europe, where 4000 km (about 2485.48 mi) would take you from Spain to Russia. Weather, economics and regulation across borders create more challenges for long-haul autonomous trucking, which may delay adoption and availability in this region and others including South America and Africa.
Why is autonomous driving happening today and not 10 years ago or 10 years from now?
There are a lot of technological advancements working together to make autonomous driving a possibility now that were not available 10 years ago. The computing power needed for autonomous driving, for example, is incredibly high. We are now at a point where an elevated level of computing power is obtainable. Furthermore, we are finally getting a good idea of the capabilities that AI and machine learning can solve across a multitude of topics – we are now discovering what is necessary to solve a research problem and how to create this type of safety-focused product.
Another crucial factor for us is the entire long-haul simulation. What we can do with simulations today is unreal; through video captures analyzed by AI we can create real-life scenarios to train and improve our systems virtually. All of these technological advancements are coming together at the same time to make autonomous driving possible.
You seem to be truly passionate about these advancements; where does that passion come from?
I am a people person. At the end of the day, these cutting-edge, future-shifting technologies, are made possible through the incredible people creating them. It is about inspiring engineers to go in the right direction, about pushing things to the limit, and about-facing complex challenges. This industry is brimming with young and talented people ready to face these challenges, and being a part of such a dynamic environment fuels my passion for Torc and the industry.
