Toyota recently invested $ 400 million in Pony.ai.
Not long ago I have just invested in Pony.ai (Xiao Ma Zhixing) $ 400 million Toyota , Today also announced a strategic cooperation with another domestic self-driving startup Momenta: Momenta will provide Toyota with high-precision map and update services based on camera vision technology, and promote Toyota ’s automated map platform (AMP) in China Landing.
Momenta was established in 2016. The company’s positioning is to develop the “autonomous driving brain”. The core technology is environment learning based on deep learning, high-precision maps and driving decision-making algorithms. The strategy is mass production of autonomous driving (Mpilot) and completely unmanned Driving (MSD) walks on two legs, the former brings data flow and cash flow to the latter, and the latter feeds back the former’s technology flow.
According to the information provided by Momenta, its high-precision maps are based on vision solutions, combined with deep learning perception technology and visual mapping and positioning technology (SLAM), using only consumer sensors (camera, GPS, IMU) strong>, you can realize automatic high-precision map construction, the relative accuracy can reach 10cm level. At the same time, map technology covers a wealth of road information, including not only geometric elements such as traffic signs, road poles, lane lines, traffic lights, road signs, but also road-level and lane-level semantic connection relationships and semantic information. This technology can be widely adapted to a variety of mass-produced sensors and can support large-scale, low-cost commercial applications. In addition, Momenta’s automated map production line can efficiently process big data to ensure the freshness of high-precision maps, thereby providing live, safe maps for different autonomous driving functions such as positioning, planning, and control.
Regarding the selection of mapping technical solutions, Momenta’s R & D director and head of map business “> Sun Gang once told the media at the end of 2018 that traditional charters use a technology route mainly based on lidar, which is too costly to deploy on a large scale, resulting in slower map updates and rapid large-scale verification. In contrast, the vision solution has low hardware costs, and is easy to deploy on a large scale and implement map updates.
Momenta obtained the qualification of domestic navigation electronic map production (Grade A) surveying and mapping in 2018.