Validation of Sensor Components with Data Replay Testing

• Test sensor behavior with recorded real-world data
• Synchronous replay of heterogeneous data streams (sensors and vehicle bus data)
• Extensive support of sensor interfaces (FPD-Link III, GMSL, MIPI CSI-2, LVDS, etc.)
• Comprehensive capabilities to monitor and analyze all replayed data in real-time  

Task

Ensuring the safety and accuracy of camera, lidar, and radar sensors under all conditions is crucial for ADAS/AD. Each sensor in the vehicle must be validated with real-world data to test its behavior in different traffic and environment scenarios. Whether for sensor development, validation, or benchmarking, a robust data replay solution is always required.

Challenge

Typically, a sensor data feed is accompanied by a bus/network data feed that includes the relevant vehicle information such as the vehicle pose. Both data feeds must be synchronized during the data replay test. In addition, the different variations of sensor interfaces means that no universal sensor interface is available. Scalability is also a challenge as scaling the test to cover thousands and millions of kilometers in the shortest time possible is crucial to achieve fast time to market.

Solution

The dSPACE data replay solution for sensor validation is based on a very modular approach. A combination of dedicated sensor interfaces and a real-time system ensures highly accurate data replay.

The Environment Sensor Interface Unit (ESI Unit) supports a wide range of standard sensor interfaces and is easily adaptable with new sensor interfaces. Through gPTP (generalized Precision Timing Protocol), the whole constellation of the ESI Unit, the SCALEXIO system, and the sensor component are synchronized, ensuring synchronous data feed. The SCALEXIO real-time system maintains low-jitter real-time replay of all the vehicle bus data, such as Ethernet(Some/IP), CAN, and FlexRay. Scalability is also granted, as multiple sensor components can be tested in parallel with a single data replay system and the entire system can be easily automated through open APIs. Another advantage is that the complete system can be used with minimal effort for closed-loop simulation in addition to data replay tests.

RTMaps, the development framework for multisensor applications from Intempora, supports the synchronous replay of recorded data streams within a wide range of file formats, such as MDF4, rosbags, DAT files, etc., in addition to comprehensive 2-D, 3-D visualization capabilities. This is advantageous for debugging and fault analysis. Real-time access on SCALEXIO and RTMaps is granted through XIL API, which can be used for monitoring, analysis, and manipulation of the replayed data in real time by the dSPACE ControlDesk software.