The core of the new platform is the QNX OS for Safety, an ISO 26262 certified operating system that supports all safety integrity levels from ASIL A to ASIL D and is therefore suitable for highly automated driving systems. The operating system has a modular microkernel architecture, which makes the integration of new sensor technologies and special ADAS processors easier, while providing the necessary fault tolerance systems, from which the highest level of reliability is required.
To help development teams to move forward quickly with new projects, the platform includes pre-integrated reference implementations for creating systems for machine vision with multiple cameras and V2X applications. The implementation of machine vision is based on the Multi-Camera Framework platform. It simplifies system design, by taking over the complex details of camera control for applications. Implementation of V2X uses the platform’s interface for remote V2X to process V2X signals from other vehicles and infrastructure elements such as traffic lights.
Additional features of the QNX Platform for ADAS include: A High Availability Manager that ensures faster recovery of software errors, POSIX APIs simplify the migration of existing ADAS applications, short boot times enable “instant-on” systems, time partitioning guarantees critical software processes the required CPU cycles, there are programming interfaces to simplify the integration of AUTOSAR additional features from third-parties as well as secure network protocols and file systems.
“The automotive industry is at a turning point. The trend towards more highly automated vehicles is moving at an ever increasing pace. However, developing software for automated driving is an extremely challenging task”, said Ian Riches, Director Global Automotive Practice at the British market research company Strategy Analytics, commenting on the development environment for software vendors such as QNX.
Complementary technologies from the partner network
Hardware and software vendors are already supporting the QNX Platform for ADAS today with SoCs, V2X modules, algorithms for machine vision, communication middleware, and other complementary technologies. The following partner technologies will be pre-integrated into the first release of the platform: SoCs, such as the Intel Atom C2000 product family, TI’s TDA2x processor family, ADAS platforms such as Nvidia Drive Automotive Platform for machine vision, deep learning and the fusion of sensor data, ADAS image processing, such as Itseez algorithms for road recognition, pedestrian detection, traffic sign recognition and forward collision warnings; TI Vision Libraries for front cameras, surround vision systems, sensor data fusion and smart rear-view cameras and V2X, for example V2X MK5-on-board units from Cohda Wireless, software stacks and applications with remote interface for data analysis and simulated GPS for simulation of numerous vehicles.