Technology is continuing to evolve, further allow humans to better interact with the world around them and increase productivity, safety, and sustainability. As technology has grown, the notion of the self-driving car or automated fleet ha really come to light. Companies such as Alphabet Inc., Uber, and General Motors, to name a few, have really gotten out of the gate hot with their concept and innovation. While they are making noticeable progress, these companies have been struggling with sensing conditions, meaning they are having difficult incorporating technology that allows for effective breaking on highways or high-speed scenarios. With that problem in mind, TuSimple has stepped in to provide a solution.
TuSimple is an autonomous tech startup, based out of San Diego, California, which is devoted towards applying artificial intelligence and computer vision technologies. The company, which operates in both the US and China, has created a system which allows for trucks to “see” vehicles on the highway more than 1,000 meters (more than half a mile) ahead of them.
Compared to competitors, the 1000 meters of visibility is a clear winner. Waymo, a function of Alphabet Inc. was able to achieve success in the monitoring system, but only a distance of 300 meters. At this distance, vehicles are able to readily track objects, and most are able to stop within that distance at high speeds. The problem with that system is that it would not be capable of stopping a Class-8 semi, which with a full load can weigh nearly 80,000 pounds. At the 300-meter distance, Waymo says they are confident that a minivan, weighing in at 5,000 pounds, would be able to stop.
The system that allows for 1000 meters of visibility utilizes a specialized algorithm and several cameras, with some focused on the short-range view and others with a long-term view. The algorithm is used to take what is captured by the cameras and “use those cues” to makes decisions. Xiaodi Hou, the chief technology officer and US unit president for TuSimple, talked more what their system identifies and understands at 1000 meters away, saying “we safely know what lane that vehicle is in, which is very important. Is it an adjacent lane or the same lane? Second, we know is that vehicle stopped or is that vehicle moving relative to you? We also know what classification of vehicle, a truck, a car, a pickup truck.”
From a purely factual standpoint, a large semi-truck could likely easily stop over 500 meters if a still object was sensed at that distance out. However, what TuSimple offers is the additional coverage, which makes to possible for the vehicle to operate at better fuel efficiency. With the extra distance, semis would have less of a need to accelerate or decelerate. To the everyday driver, increased efficiency in fuel usage may not be the primary concern, but in the trucking industry, fuel is the single highest expense
TuSimple was able to complete the system in large part due to its $83 million in funding, including large contributions from Nvidia.