Volvo combines crash data history with AI simulation to redefine vehicle safety

Volvo’s software-defined approach combines decades of safety innovation with cutting-edge simulation technology to create safer vehicles. The Swedish automaker is leveraging its extensive crash data history to develop new safety features while using advanced “Gaussian splatting” to virtually test cars in thousands of scenarios derived from real-world conditions. This technological approach reflects Volvo’s longstanding commitment to prioritizing passenger safety over performance metrics.

The big picture: Volvo is embracing the software-defined vehicle (SDV) architecture, replacing numerous electronic control units with powerful central computers that manage key vehicle systems.

• The company organizes these computers around specific domains including infotainment, advanced driver assistance, handling and powertrain, and interior comfort.
• This centralized approach enables more sophisticated safety features and testing capabilities compared to traditional vehicle architectures.

Decades of data collection: Volvo’s safety innovations stem from extensive real-world accident research dating back to the 1970s.

• “We’ve been visiting crash sites since the ’70s. We’ve been recording event data from a fleet of cars for many, many years,” explained Alwin Bakkenes, head of global software engineering at Volvo Cars.
• This accumulated data has driven the development of numerous safety features, from the three-point seatbelt to the latest lidar-based protection systems.

Breakthrough simulation technology: Volvo’s implementation of “Gaussian splatting” represents a significant advancement in vehicle safety testing.

• The technology can transform a single traffic scenario into “thousands or tens of thousands of scenarios from real-world data,” according to Erik Coelingh, VP of product at Zenseact.
• By using neural networks to reconstruct scenes from sensor data, Volvo can virtually test vehicles across a vastly expanded range of conditions than would be possible with physical testing alone.

Latest innovation: Volvo is introducing “lidar AS” in its ES90 model, a system designed to protect vulnerable road users.

• The feature uses lidar technology to detect and steer away from pedestrians and other vulnerable road users, particularly in low-light conditions.
• This system builds upon Volvo’s history of safety innovations while leveraging their modern software-defined architecture.

Why this matters: While physical testing remains essential, virtual simulation dramatically expands Volvo’s ability to identify safety issues and test solutions before vehicles reach production.

• The combination of historical crash data with advanced simulation creates a powerful platform for improving vehicle safety across a comprehensive range of scenarios.
• This approach allows Volvo to maintain its brand identity as a safety leader while modernizing its development processes for software-defined vehicles.