Sweden’s Arriver is developing software for perception, drive policy, and everything in between, for the next generation of cars. Veoneer’s dedicated software unit recently announced plans to join forces with BMW and Qualcomm Technologies, Inc. to co-develop Automated Driving Systems.
Giuseppe Rosso, Arriver’s President and Managing Director, has been talking with Auto Futures.
Arriver was created in January, 2021, as part of a collaboration between Veoneer and Qualcomm. Veoneer has subsequently been fully acquired by Qualcomm.
Arriver’s goal is to be one of the key players in the market for Advanced Driver Assistance Systems (ADAS), collaborative and autonomous driving software.
“It builds on more than a decade of experience in Active Safety software development and delivers an open, scalable and flexible architecture solution running today in production with actual product generations and on Qualcomm® Snapdragon Ride™ System on a Chip (SoC) platform for sop 2024/25. Our goal is to provide leading software stacks enabling supervised driver assistance systems to unsupervised highly automated driving,” says Rosso.
Arriver’s Vision Perception software uses the latest artificial intelligence and deep learning techniques to offer functions such three-dimensional object detection, path prediction and probe data map building.
“Arriver’s Vision Perception software is in production today and evaluated against NCAP tests with some of the highest scores, indicating world-class feature solutions and performance.”
The company is collaborating with Qualcomm to deliver scalable ADAS and highly automated driving solutions.
“This platform will also address the growing needs of the automotive ecosystem for scalable and upgradable solutions, which requires highly advanced and power-efficient compute, connectivity, maps and cloud service capabilities across all vehicle tiers. This strategic collaboration effectively enables Qualcomm as a go-to market to Tier 1 suppliers and OEMs for Arriver software,” he explains.
Rosso says Arriver’s key strength is its ability to be open, scalable and flexible.
“As a developer of core software for automation, we must be able to be open to integrate other technology companies’ solutions to our core to a varying degree, a closed system will eventually run into a dead end, will not be able to reach best in class performances and limit the business opportunities ahead. This encompasses every aspect of what makes up a car, from how it moves (horsepower, acceleration), to looks, electrification, and automation.
“We are flexible as auto manufacturers will have different needs and requirements, also down to the individual vehicle model level.
“We want to create compatibility between as many vehicle tiers as possible, from the more basic mass-market vehicles to the cutting-edge of highly automated premium segment vehicles. Doing so requires scalability of both software and hardware, including the system-on-a-chip (SoC) platform,” says Rosso.
Collaborating With BMW
In November 2021 Arriver, BMW and Qualcomm announced a long-term development cooperation for the development of Automated Driving technologies. BMW’s next generation automated driving system will be based on the Snapdragon Ride Vision system-on-chip (SoC) including Arriver Computer Vision.
The cooperation will see specialists working together in locations worldwide, including Germany, US, Sweden, China, Romania, and the BMW AD Test Center in Czech Republic.
“BMW’s current AD Stack with Arriver’s NCAP Stack combines to create a team of 1400 experts having the best competence of BMW, Qualcomm and Arriver. The platform, based on Qualcomm’s System-on-Chip and Arriver’s Vision Perception, will establish best-in-class NCAP to L2/L3 Automated Driving functions,” says Rosso.
In a press release, Nicolai Martin, Senior Vice President Driving Experience, BMW Group, commented: “To enable sophisticated and safe functionalities in a vehicle you need state-of-the art software in all components of the digital value chain. This forms the backbone for intelligent driver assistance systems. The BMW Group is excited to further extend our partnership with global technology leaders Qualcomm Technologies and Arriver to include a long-lasting strategic co-development cooperation in order to continue delivering world-class driving experiences to our customers.”
Nakul Duggal, Senior Vice President and General Manager Automotive, Qualcomm Technologies, Inc. added: “This cooperation will position us to extend BMW Automated Drive software to the Snapdragon Ride platform, and expand the accessibility of safer self-driving technologies to other automakers and Tier-1s in an open, flexible and scalable deployment framework.”
“We see a safer world with more sensors and safety software in vehicles with human drivers.”
Rosso strongly believes that industry collaborations are the key to achieving the goal of safer driving across Europe.
“Development of unsupervised Driver Assist software is not an undertaking a single technology company can do without a fleet of vehicles feeding development teams with data from the real world for development and pre-deployment. Collaboration enables access to data mandatory for proper feedback and development to continue meeting
and exceeding road safety standards and customer expectations for years to come.”
“Additionally, investments for the growth of supervised ADAS vs unsupervised are too big for a single player OEM or supplier. To achieve future-proof performance, a vertical collaboration is needed. Consolidation of ECU is changing the business model from the traditional tier 1 OEM and, to follow the new one, collaborations are becoming a must.” he says.
He is excited at the role AI will play in making our roads a safer place to drive and travel in the future.
“The growth of AI and all of the possibilities. Really focusing efforts using data onboard (sensing and control) the vehicle as well as offboard (cloud, maps). Cross domain consolidation and simplification of overall systems.”
As for driverless transport, Rosso has this to say: “Currently, Mercedes Benz was approved for L3 autonomy in Germany using our generation four vision perception stack which has a limited operational design domain. What we see is the expansion of that operational design domain based on the availability of new sensors and new data along next 2-3 years.”
Finally we asked Rosso for his thoughts on what urban mobility look like by 2030.
“We see a safer world with more sensors and safety software in vehicles with human drivers. In addition, it is likely that some autonomous taxi services will be operating in large cities within a specific operational design domain,” he concludes.
Keyword: “Collaborations are Becoming a Must” for Safer Driving – Arriver President Giuseppe Rosso