Top > News & Events > Visual Perception for Autonomous Driving

News & Events

Visual Perception for Autonomous Driving


ADLINK and TIER IV provide hardware and software integrated turn-key solutions based on Autoware* and ROS 1/2 for visual perception in autonomous driving applications.

The integration of TIER IV C1/C2 Automotive HDR Camera and ADLINK RQX-58G AI-based Perception ECU has been pre-tested, pre-verified and pre-configured, allowing users to develop their visual perception solution without any complex processes, and ensuring accurate imaging results for autonomous driving applications.

Sensor Fusion & Synchronization
With the integration of TIER IV's C1/C2 cameras and ADLINK's RQX-58G AI-based Perception ECU, cameras can be trigged and controlled separately and the image data from multiple cameras can be perfectly synchronized with LiDAR data to achieve sensor fusion. Not only can the solution be used for image recognition in perception tasks, but it can also compress and transmit image data for remote monitoring.

TIER IV Automotive HDR Camera (C1/C2)
~120dB HDR, motion artifact mitigation and LED flicker mitigation (LFM)
~Integrated ISP for correction of imaging and lens distortion at the edge to reduce ECU computation load
~Flexible resolution options: C1 2.5Mpix, C2 5.4MPix
~Variety of lens options: 46° to 180° angle of view supported for percep tion and remote monitoring
~Autoware plug-and-play (ROS 1/2 driver support)
~Operating temperature range -40 to 85 ℃
~IP69K dust/water-proof and ingress protection
~5-axis active alignment process to adjust lens into proper alignment
~Gigabit Multimedia Serial Link 2 (GMSL2) supports automotive multistreaming and HW/SW triggering
~Compliant with ISO 26262 functional safety level ASIL B* and IATF 16949 automotive quality management system
~High reliability with AEC-Q100, Q101 and Q200 certificated parts

ADLINK Robotic Controller / AI-based Perception ECU (RQX-58G)
~Powered by NVIDIA® Jetson AGX Xavier™ module
~Excellent performance per watt with power consumption as low as 30W
~Comprehensive I/O for connecting a wide range of sensors such as camera, LiDAR, GNSS and IMU
~Integrated ADLINK NeuronSDK™ to simplify and streamline development; built-in DDS ensures reliability with shared memory to reduce latency
~Supports up to 8 cameras via GMSL 2 connection to provide high-bandwidth and high-speed image data transmission
~Accurate time synchronization of data from various sensors (video cameras, LiDAR, GNSS) for sensor fusion
~512-core Volta GPU with Tensor Cores for AI image recognition
~Pre-installed TIER IV C1 camera driver for simplified integration