For a better experience on dSPACE.com, enable JavaScript in your browser. Thank you!

MicroAutoBox Embedded SPU

Compact and robust in-vehicle prototyping unit for multisensor applications

The MicroAutoBox Embedded SPU is a compact and robust prototyping unit with high processing power and an intuitive software environment for the development of advanced driver assistance systems and highly automated driving functions. It can be used as a stand-alone system or as an extension to MicroAutoBox II.

 

Please note:
MicroAutoBox Embedded SPU is still under development and planned to be available in the third quarter of 2017. Until then, the information provided on this page is preliminary and subject to change without notice.

Compact and robust in-vehicle prototyping unit for multisensor applications

Multisensor applications play an essential role in many areas, such as advanced driver assistance systems (ADAS), automated driving, self-driving vehicles, and robotics. The development of these kinds of applications requires a dedicated hard- and software environment that can process and fuse data from various sensors, such as cameras, lidars, radars and GNSS receivers, calculate motion control 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 enormous amount of data, 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 (Real-Time Multisensor applications) is precisely tailored to this use case.

MicroAutoBox Embedded SPU is a unique combination of high processing power, interfaces to automotive vehicle networks, environment sensors, GNSS positioning, wireless communication, and an extremely compact and robust design for in-vehicle use. Using RTMaps, you can directly access and configure all interfaces of the Embedded SPU by means of ready-to-use block diagrams and design your algorithms for sensor data processing, sensor fusion, etc. in short iteration cycles and in a graphical environment. Each sensor data sample is captured along with its time stamp at its own genuine pace. This ensures that all data is time-correlated. It can even be recorded to an external mass data logging extension with high data throughput and then be played back later on. In addition, it is possible to process algorithms on multiple distributed Embedded SPUs while preserving time coherency and synchronization of heterogeneous data streams.

Key Benefits

  • Stand-alone system or extension to MicroAutoBox II
  • Multicore ARM® CPU with embedded GPU from NVIDIA®
  • Interfaces for cameras, BroadR-Reach, CAN/CAN FD, LIN, Ethernet, USB, WLAN, LTE, and Bluetooth 
  • GNSS receiver with inertia measurement unit (IMU)
  • Optional mass data logging extension 

Parameter Specification
Processing Units

CPU:

  • Two NVIDIA® Denver 2 cores and four ARM® A57 cores (up to 2 GHz and 2 MB L2 cache each)

GPU:

  • 256-core NVIDIA® PascalTM at up to 1300 MHz
Memory

RAM:

  • 8 GB 128-bit LPDDR4 RAM

Flash:

  • 32 GB eMMC plus 128 GB M2 card

Optional mass data logging unit

Interfaces
  • 4 x Gigabit Ethernet, 2 x Gigabit Ethernet via internal switch
  • 2 x USB 2.0, 2 x USB 3.0
  • WLAN 802.11 n/ac, Bluetooth 4.1, optional LTE
  • 2 x HDMI 2.0 Out display interfaces
  • 4 x CAN/CAN FD, 2 x LIN (Master/Slave), 2 x BroadR-Reach
  • 4 x GMSL In, 2 x HDMI 1.4b In camera interfaces (other interfaces on request)
  • 4 x Digital In, 4 x Digital Out, 4 x Analog In
  • uBlox NEO-M8U GNSS receiver (GPS, GLONASS, Beidou, Galileo) with integrated inertia measurement unit (IMU) and support for Untethered Dead Reckoning (UDR)
Operating system
  • Linux for Tegra from NVIDIA
Software support

Graphical development environment:

  • RTMaps (Real-time Multisensor applications)

GPU programming language:

  • NVIDIA® CUDA®

Deep learning:

  • NVIDIA® TensorRTTM, cuDNN®

Computer vision:

  • NVIDIA® VisionWorksTM, OpenCV
Physical characteristics Enclosure size:
  • Approx. 200 x 225 x 50 mm (7.9 x 8.9 x 2.0 in)

Operating (case) temperature:

  • -20 ... +70 °C (-4 ... +158 °F)

Power supply:

  • 6 ... 40 V DC input power supply, protected against overvoltage and reverse polarity

Power consumption:

  • Max. 50 W
Certification Electromagnetic compatibility (EMC):
  • EN 61326-1 Table 2
  • CISPR 11, EN 55011 Group 1, Class A
  • FCC 47 CFR Part 15

Vibration:

  • ISO 16750-3:2007 / 4.1.2.4 Test IV
  • EN 60068-2-6

Shock:

  • ISO 16750-3:2007 / 4.2.2

Further Information