High Performance Sensor Data Processing

• MicroAutoBox Embedded SPU¹⁾ with Multicore ARM^® CPU and NVIDIA^® GPU for prototyping multisensor applications in the vehicle
• Various interfaces to environment sensors, such as camera, radar and lidar, GNSS positioning, and wireless communication
• RTMaps Developer version on standard PCs to develop algorithms in C/C++, Python, NVIDIA CUDA or Open Computing Language (OpenCL)
• RTMaps Runtime version to execute algorithms on MicroAutoBox Embedded SPU

High-performance sensor data processing is essential for automated driving and self-driving vehicles. Developing these kinds of applications requires a dedicated hardware and software environment that can process and merge data from various sensors, such as cameras, radars, lidars, and GNSS receivers, calculate application algorithms, and connect to actuators or HMIs. In addition, sensor and vehicle network data has to be recorded and played back time-synchronously for testing purposes.

To prototype associated algorithms in the vehicle and to process the large data volumes, a compact and robust prototyping unit with high processing power and an intuitive software development environment is required. The MicroAutoBox Embedded SPU in combination with RTMaps is perfectly tailored to this use case.

¹⁾ MicroAutoBox Embedded SPU is still under development. The information about MicroAutoBox Embedded SPU is subject to change without notice. For more information, please contact dSPACE.