Technology

The race to autonomy is a marathon, not a sprint.

Our Approach

Autonomous vehicles have the potential to change our world. They can provide accessible mobility for more people, transport goods efficiently and help make our roads safer. We are working to make this a reality by creating commercialization-ready autonomous solutions, proven safe and carefully crafted to meet real industry goals.

Our approach to developing technology is iterative. Safety dictates our timeline for commercialization and customer value is our driving priority.

Building the virtual driver

We designed our technology to emulate the driving styles of highly skilled and safe drivers. Our system bases its driving decisions on an approach we call See-Think-Act. The system first perceives the world and understands its location (perceiving and localizing). It uses these factors to calculate the optimal driving decision (thinking). Finally, it executes this decision in real-time to navigate its environment. Our system continuously performs these tasks as it traverses the roadway.

Autonomous truck sensor graphic

See: A Full Sensor Suite

Our current system uses integrated cameras, long and short-range Lidar, and long and short-range radar to detect its surroundings. This sensor suite utilizes each component’s strengths to provide a 3D picture of the environment, detect objects and calculate the movement of vehicles around the truck for better prediction and decision-making.

Think: Planning The Next Move

Road environments and traffic conditions can change in the blink of an eye. Torc vehicles use detailed HD maps to plot out travel routes and ensure the safest and fastest trip possible. Our advanced localization and prediction algorithms allow our vehicles to immediately determine the next safe move on busy and unpredictable roadways by planning and updating routes in real-time.

Act: Safe Maneuvering

Level 4 behaviors respond to complex driving situations such as lane changes, merges and sudden speed variations in traffic. These abilities are what transforms driver assist technology into true Level 4 autonomy.

Testing

Creating an autonomous solution requires a building block approach where each stage is thoroughly understood, tested and verified before transitioning to the next evolutionary step.

Software Validation

Rapid innovation requires following a build-test-learn cycle that allows you to quickly test technological advancements in the real world. We’ve used this concept for a decade to refine and expand our self-driving technologies in every environment imaginable — from off-road terrain to highways to congested city streets. We believe in controlling the full software stack. We always use raw, unprocessed sensor data and a mix of sensor technologies, including lidar, radar, vision (cameras), etc. Software and hardware needs cyclically drive each other as we develop and test new systems. And we’re continually researching and integrating emerging technologies such as deep and reinforced learning to help make our systems smarter, faster and more reliable.

torc team member checking maps.

From Closed Course to Public Roads

As part of Torc's comprehensive safety process, we test our trucks extensively on closed course tracks. This testing is critical to successfully advancing automated technology. After months of testing and safety validation, our vehicles' deployment on public roads takes place on a closed-loop track. Learn more about how Torc is developing autonomous solutions for Trucking.

Trucking
An autonomous truck driving over a bridge.