China’s auto market is the world’s largest and fastest growing. In a highly competitive environment, local manufacturers are forced to woo pampered customers with advanced driver assistance features, and Western component suppliers are not standing aside. Qualcomm recently agreed to partner with China’s DeepRoute.ai.
Image source: Qualcomm Technologies
The latter is a startup founded in 2019, which has already managed to attract more than $500 million in capital and develop the DeepRoute IO autopilot platform, which allows real-time analysis of the road situation, including without the help of lidar, and also does without detailed digital maps of the area.
In collaboration with Qualcomm, DeepRoute.ai intends to create active driver assistance systems based on Snapdragon Ride hardware solutions. The list of solutions offered by the partners will include both flagship solutions using lidars and more economical ones based only on images received from cameras. According to DeepRoute.ai, it has achieved particular success in assessing the surrounding environment in both very bright and insufficient lighting. This allows it to achieve the same accuracy of object recognition using cameras alone as when using lidars. The platform can simultaneously process information from 16 cameras, several radars and lidars, preventing collisions with pedestrians, ensuring lane keeping and obstacle recognition in real time. DeepRoute.ai solutions are suitable for both electric cars and internal combustion engine vehicles.
Based on the joint platform with Qualcomm, solutions of different levels will be offered, providing automatic transport control both on suburban highways and in urban conditions, as well as automatic parking and other service functions. Cooperation with Qualcomm will allow the Chinese DeepRoute.ai to offer its solutions to customers around the world. It is difficult to say to what extent the content of the press release was influenced by the incident with the Xiaomi SU7 electric car that crashed in China, but the text mentions an example with road barriers placed along the route. According to representatives of DeepRoute.ai, its end-to-end AI models allow timely rebuilding upon detection of such obstacles on the road, avoiding a collision in automatic mode.