Qualcomm & Alps Alpine Announce ‘ViewPose' a Advanced Camera-based Vehicle Positioning Technology

author: Eric Walz   

Autonomous vehicles, as well as conventional vehicles equipped with GPS navigation or advanced driver assist systems (ADAS), rely upon the vehicle's precise position to safely navigate and keep occupants safe. In the field of autonomous driving and robotics, a vehicle's precise location and orientation is known as "pose" and its required before a self-driving vehicle can begin to navigate using maps.

This week, Japanese electronics company Alps Alpine Co Ltd. and Chipmaker Qualcomm announced a new more accurate positioning technology called "ViewPose" that helps address the drawbacks of conventional GPS, especially in urban areas with spotty signals. 

Alps Alpine developed a new camera-based sensing and positioning system that supports "absolute lane-level vehicle positioning" in virtually all environments and its supported by Qualcomm's technology.

ViewPose uses Qualcomm's Snapdragon Automotive 5G platform, which supports Multi Frequency Global Navigation Satellite System (MF-GNSS), combined with the 3rd Generation Qualcomm Snapdragon Automotive Cockpit Platform for the processing and stitching of multiple camera images to determine a vehicle's precise location.

Qualcomm's Automotive Cockpit Platform is supported by the company's new Vision Enhanced Precise Positioning (VEPP) software. ViewPose is one of the first camera-sensing devices to feature the Qualcomm VEPP software.

Millions of drivers rely upon the world's Global Positioning Systems (GPS) for turn-by-turn directions. Cell phones that run navigation apps like Google Maps or Waze have an accuracy of around 4.9 meters (16 ft). While this level of accuracy is precise enough to get to where you're going, a self-driving vehicle requires more precise positioning data to safely navigate, especially in urban areas with unreliable GPS signals.

The VEPP software combines the output of multiple automotive sensors, including a Global Navigation Satellite Services (GNSS) receiver, a forward-facing camera, Inertial Measurement Unit (IMU), and wheel sensor data to offer accurate global vehicle positioning accurate to less than 1 meter. The wheel data includes steering angle and speed from a vehicle's CAN bus network. The data is collected via a sensor board mounted in the vehicle. It's small enough to be mounted above the rearview mirror on the windshield. 

Qualcomm said that VEPP does not rely on the availability of any kind of HD map, like most autonomous vehicles require for navigation.

"Alternate solutions, such as using Lidar map scan, achieve highly accurate positioning and attitude in most environments, however these approaches are costly and limited only to areas where Lidar point cloud maps are available," said Nakul Duggal, senior vice president & GM, automotive, Qualcomm Technologies, Inc. " Working with Alps Alpine, we will provide a globally accurate lane-level positioning solution for challenging environments by leveraging already available sensor information in the car for a fraction of the cost of substitute technologies."

View Pose addresses the problem of poor GPS reception in dense urban areas due to tall buildings and other infrastructure. It can even maintain accurate vehicle pose with poor GPS reception, such as in tunnels and covered parking garages, the companies said.

For developers of autonomous vehicles, Qualcomm's ViewPose software can aid in matching lanes, signs or landmarks to high definition maps used for navigation. ViewPose can also assist in the creation of crowdsourced HD maps using the camera and various sensors to create a map based on real-world coordinates.

By leveraging MF-GNSS information provided by the Snapdragon Automotive 5G platforms, the ViewPose positioning software is precise enough to identify if a vehicle is traveling in an exit-only lane, or a turn-only lane and can be used to prompt the driver to change lanes. It can also assist in navigation, such as letting a driver know how many lanes they need to cross to reach their exit on a multilane highway.

ViewPose offers an update rate as high as the camera frame rate, which can range from 10 to 60 frames per second. 

With new connected technology being developed, including cellular vehicle-to-everything (C-V2X), which allows vehicles to communicate with infrastructure, traffic signals or pedestrians, automakers are looking for more accurate position-location technology to increase safety for all road users.

However, systems such as high-precision GPS can cost tens of thousands of dollars, according to Qualcomm, which prevents them from being adopted for wide-scale deployments in the auto industry. ViewPose addresses this need with a low-cost, yet highly reliable alternative.

With this year's annual CES show cancelled due to the global pandemic, the technology was announced as part of Qualcomm's Technology showcase, a virtual event  held on Jan 26-27 to highlight Qualcomm's new products for the automotive industry that otherwise would have been announced at CES.

Alps Alpine, which is a leading manufacturer of electronic components and automotive infotainment systems. The company developed the technology with Qualcomm to provide a highly accurate vehicle positioning in virtually all driving environments. 

ViewPose is currently in development targeting OEM deployment starting 2024. 

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.