Post by AUTOPIA Program

🚗🔬 Advancing Software-Defined Vehicles through Hardware Abstraction Within the framework of the European project HAL4SDV (Hardware Abstraction Layer for a European Software-Defined Vehicle), Programa AUTOPIA is developing technologies that enable perception software to operate independently of the underlying sensor hardware. One of these developments is our LiDAR Abstraction Layer, designed to remove dependencies on vendor-specific drivers by providing a common interface that makes different LiDAR sensors interchangeable through plugins. 🔹 Vendor-specific details are isolated in dedicated plugins 🔹 A common abstraction layer provides unified access to LiDAR data 🔹 Applications can interact with different sensors through the same API 🔹 New LiDAR models can be integrated with minimal impact on the rest of the software stack. The architecture has been validated on our autonomous driving test platform, equipped with multiple LiDARs, cameras, GPS, and a dual-computing platform for perception and decision-making. By feeding the AUTOPIA Perception Stack, the abstraction layer enables dynamic occupancy grid generation and real-time object tracking, while maintaining full compatibility across different sensor manufacturers. 🎯 The result? Seamless hardware interchangeability. In the demonstration below, the same perception pipeline runs with LiDARs from different vendors - RoboSense and Hesai Technology - without any modifications to the upper software layers. This work aligns with HAL4SDV's vision of creating standardized hardware abstraction mechanisms that decouple vehicle software from hardware platforms, paving the way for the next generation of Software-Defined Vehicles (SDVs). CAR (CSIC-UPM) Jorge Villagra Serrano | Diego Aceituno Seoane #HAL4SDV #AUTOPIA #CSIC #SoftwareDefinedVehicle #SDV #AutonomousDriving #LiDAR #Perception #Robotics #EmbeddedSystems #AutomotiveSoftware #SmartMobility #Innovation #HorizonEurope #ChipsJU

Video Content