Hardware-in-the-Loop Injection Testing for Autonomous Driving Leveraging NVIDIA ORIN, FPGA, and AI

Solution Overview

In the field of autonomous driving, validating the algorithms of domain controllers is a critical step to ensure that autonomous driving technology can be safely and reliably deployed to the market.

Hardware-in-the-loop (HIL) injection testing is one of the important testing methods in the development process of autonomous driving systems. By means of simulation or data playback, it tests the performance of autonomous driving systems in various complex scenarios, maximizing the coverage of laboratory tests, reducing the number of field road test items, and improving R&D efficiency.

Addressing the challenges of autonomous driving domain controller algorithm validation, Sienovo introduces a hardware-in-the-loop injection testing solution, which covers hardware, software, and customized services, provides a unified toolchain, meets multi-scenario testing needs in a one-stop manner, significantly reduces testing costs, and helps improve R&D efficiency.

Solution Advantages

Millisecond-Level Synchronization Accuracy

Achieves high-precision time synchronization between multi-channel camera data, other sensor data, and bus data, ensuring that the test environment can accurately simulate actual system operation.

Time synchronization accuracy less than 1ms

Fast Real-time Response

Optimizes the data transmission link, achieves instant response, ensuring that the link delay for image injection into the ECU is within 1 frame.

Injection delay less than 1 frame

High Data Consistency

Provides various types of data playback, achieves high-precision synchronous injection, ensuring data real-time performance and consistency in data content, frame rate, and timing.

Ensures I/O data content consistency

Features dynamic buffer design to ensure stable frame rates and no frame loss

Controls data transmission timing via precise timers to ensure time alignment

Extensive Compatibility and Expandability

Strong adaptability and extensive compatibility, capable of continuously integrating various types of sensors, meeting future testing requirements.

Supports multi-channel expansion and integration of autonomous driving sensors such as cameras, radar, and LiDAR.

Supports parsing of various video data source formats such as YUV, RAW, H264, and H265.

Compatible with various serial chips from MAXIM, TI, ROHM, and THINE.

Compatible with common sensor hardware protocols such as GMSL and FPD-LINK.

Compatible with mainstream AD domain controllers such as NVIDIA ORIN, Horizon Robotics, and TI TDA4.