DeepRoute Announces its New Full-Stack Autonomous Driving Sensing Solution

author: Eric Walz   

DeepRoute, a AI-based startup developing a level 4 level self-driving full-stack solution, announced yesterday the release of its new sensing solution called DeepRoute-Sense, which is a all-in-one self-driving system capable of fusing data from multiple sensors, enabling autonomous vehicles to better perceive their surroundings. It also an easier way for OEM's to deploy driverless technology.

DeepRoute-Sense aims to enable developers of autonomous driving technology, including OEM's, to achieve quicker sensor deployment with an efficient and reliable perception system. The perception system helps a self-driving car to safely navigate its environment. 

DeepRoute is based in Shenzhen, Beijing and also operates a research center in Silicon Valley. The company is focused on developing Level 4 autonomous driving technology. Level 4 autonomy means that no human intervention is required under most driving conditions. 

DeepRoute is one of this year's CES 2020 Innovation Award honorees in the category of Vehicle Intelligence & Transportation, which recognizes outstanding design and engineering in technology products. 

DeepRoute's autonomous driving setup includes a small set-top box that's mounted on the roof of the vehicle. The set top box includes hardware for robust sensor fusion. The entire package is lightweight and compact, measuring just 31 centimeters tall.

The roof box includes 8 vehicle cameras, millimeter wave radar, 3 LiDARs, Global Navigation Satellite System (GNSS) and a series of other sensors to help correspondence and data synchronization between the controllers. 

Inside the set-top box, there is also a dedicated sensor data processing device called ADS Synchronous Controller. The product pre-processed and fuses the massive amounts of sensor data collected from the vehicle's cameras, lidars, GNSS and other sensors through high-precision time and space synchronization. 

The ADS Synchronous Controller also supports deep neural networks (DNN) and simultaneous location and mapping (SLAM) hardware acceleration, which greatly increases computational efficiency, DeepRoute says.

DeepRoute designed the mechanical structure of the roof box in-house. The set-top box has lighter weight, smaller volume and higher impact resistance to other products currently on the market according to DeepRoute. The roof box can be easily mounted to the roof of different self-driving vehicle models for sensor deployment.

DeepRoute's set-top box is compact. (Photo: DeepRoute)

DeepRoute Sense plans the optimal route according to map and real-time traffic data as well as planning the optimal path according to the vehicle's surroundings. The system also has multiple hardware and software safeguard mechanisms, which assure the safety and stability of the entire system.

DeepRoute also can generate HD maps for autonomous driving by fusing the data collected from its set top box with the traditional maps and can label road lanes automatically for efficient map creation.

"The team at DeepRoute has worked hard on the development of DeepRoute-Sense and we are excited to finally share our technology with the industry," stated Shuang Gao, COO of DeepRoute. "By bringing this to the market, we are hoping to fuel progress within the industry and bring full vehicle autonomy to be one step closer to the masses." 

One of the hardware highlights is an in-vehicle camera created by DeepRoute. The camera features anti-glare, anti-ghosting and highly reliable signal detection. The camera costs less than other other industrial cameras while offering a higher dynamic range. 

DeepRoute's camera was specifically developed to withstand extreme weather conditions and can be used during intense sunlight, or when another vehicle's high beam headlights shine directly into it, without overexposing the image due to too much light.

DeepRoute's employees come from leading companies, including Google, Microsoft, Baidu and Ford. The team at DeepRoute has experience in various fields of autonomous driving, HD mapping and positioning, planning and control, hardware systems, infrastructure, computer simulation, cloud computing and storage.

DeepRoute has also developed a data platform which supports the operations of fleets with 100 or more vehicles.

In September, the company it raised $50 million in pre-Series A funding led by Fosun RZ Capital, the venture capital arm of the Chinese conglomerate Fosun International.

In October, the company was granted a permit by the California DMV to test its self-driving vehicles on public roads in the state. DeepRoute joins around 60 other companies granted a permit in California.

DeepRoute's initial fleet of autonomous vehicles in California will help to refine the company's sensor fusion technology.

Eric Walz

Originally hailing from New Jersey, Eric is a automotive & technology reporter covering the high-tech industry here in Silicon Valley. He has over 15 years of automotive experience and a bachelors degree in computer science. These skills, combined with technical writing and news reporting, allows him to fully understand and identify new and innovative technologies in the auto industry and beyond. He has worked at Uber on self-driving cars and as a technical writer, helping people to understand and work with technology.