>

ASEva data acquisition and analysis platform

ASEva system is Zeer's self-developed, lightweight, highly reliable and customizable integrated platform for acquisition and analysis. The system has a complete set of scene data acquisition and processing tool chains for intelligent driving scene data acquisition and extraction, and provides tools for converting real scene data to virtual simulation data, thus allowing real scenes to be accurately reproduced in virtual environments. Meanwhile, The system can also be applied to functional safety and SOTIF, ground truth systems, driving behavior analysis, functional evaluation analysis, etc.

ASEva system has a powerful local and cloud architecture. Data from the local side can be uploaded to the cloud in real time. The cloud data platform can realize multi-user management, multi-vehicle and equipment cloud monitoring, cloud task scheduling, remote data retransmission and playback, cloud big data management and processing, tagging and interception. Different hardware and software program configurations can be customized according to users' actual needs, providing users with a full set of solutions that better fit their needs.ASEva currently has a market share of over 80% and serves customers mainly from major OEMs, component suppliers and testing and certification bodies across China.


ASEva EN.png


开放道路2.jpg


开放道路3.jpg


开放道路4.jpg


Interface display

开放道路5.jpg

开放道路6.jpg




Application Profile

1、Scenario data acquisition and extraction

ASEva data acquisition system can provide real road dynamic traffic flow and static traffic information acquisition, and provide data post-processing platform. ASEva system can be used for scene annotation, extraction, slicing and massive scene management.

  • Compatible with a variety of scene extraction methods: support for manual scene extraction, automatic scene extraction, support for scene signal primitive extraction, support for user-defined scene extraction.

  • Built-in self-extracting templates for rich scenarios, such as vehicle cut-in, cut-out, overtaking, lane change, etc.

  • Data annotation, which can be used to train perception.

  • Can be used for scenario-based autonomous driving scenario library construction to support scenario-based autonomous driving function development and verification testing;Can be used for driving behavior library construction to assist customers in analyzing the differences between manual driving and autonomous driving to help customers improve autonomous driving performance.

  • Provide secondary development interface, which can assist customers to establish rapid scene extraction capability.


开放道路7.jpg



2、GT Test System

ASEva Ground Truth System provides two types of truth systems: a multi-sensor fusion-based truth system, which can be used to develop open road testing and sensor alignment; and a V2X-based solution, which relies on high precision positioning and MESH networking technology to achieve real-time distance calculation between limited targets, such as between vehicles, between people and vehicles, and between vehicles and lane lines, which can be used for scenario-specific measurement and sensor alignment.

  • Providing a high precision true value system with a range error of around 6-8cm for multi-sensor fusion schemes target objects and up to 4cm or less for high precision positioning schemes.

  • It can be used for sensor alignment, automatically comparing the sensor under test with the true sensing results and outputting a sensing performance KPI report with multi-dimensional analysis such as target recognition accuracy, miss detection rate, false detection rate, etc.

  • Provide assessment secondary development interface.

  • Support for report customization.


开放道路8.jpg



3、Virtual Simulation

The ASEva simulation tool chain generates OPENDRIVE and OPENSCENARIO simulation standard formats from scenes acquired on the open road through ZEER's unique series of automated post-processing processes, and supports the direct import of data from mainstream simulation platforms such as VTD, Prescan, Cognata, Carmaker, 51World, etc.

  • Supporting the construction of virtual simulation scenario libraries for natural driving and the construction of virtual scenario libraries for standard regulations.

  • Support for target refinement, lane line refinement and scene reconstruction and secondary editing.

  • Support for scenario retrieval, filtering and statistical analysis.

  • Support for SIL, MIL, VIL, DIL testing and customized development.

开放道路-6.jpg


4、Analysis of natural driving behavior

ASEva natural driving behavior analysis system is based on natural driving scenarios on real open roads, driver driving characteristics for driving behavior analysis, providing driving behavior characteristics and behavioral differences under different combinations of driving people, countries, cities, roads and other conditions, outputting relevant characteristic parameters and behavioral indicators. It is widely used to guide the testing and development of active safety functions, the design of autonomous driving functions and the optimization of autonomous driving performance to improve the comfort and handling of vehicles.

  • Support Cut in/out natural driving behavior analysis, lane change natural driving behavior analysis.

  • Support analysis of natural driving behavior of following vehicles, natural driving behavior analysis of uphill and downhill and hill starts.

  • Support natural driving behavior analysis at intersections, traffic lights, intersection starts, etc.; natural driving behavior analysis at toll stations driving in, driving out and natural driving behavior analysis in specific areas such as service areas, tunnels, curves, etc.


开放道路-7.jpg


5、In-field functional testing

The ASEva RTS test system is a data acquisition and analysis software for in-field vehicle performance tests, which can be used in combination with V2X technology for specific scenarios, such as.

  • E-NCAP/C-NCAP safety assessment of vehicle ADAS functions.

  • GBT 4970-1996 Test method for vehicle smoothness random input driving, GB/T 6323-2014 Test method for vehicle handling stability, GBT/T 33577-2017 Test method for forward collision warning of vehicles in intelligent transport systems, JTT 84-2014 Test method for anti-rollover stability performance of operating vehicles, GBT 26773-2011 Test method for lane departure warning.

  • JTT 1242 operating vehicle automatic emergency braking system performance test methods and other standard regulatory tests.

  • Customised assessment reports are available.

开放道路12.jpg


6、Open Road Testing

ASEva open road test system is mainly used for vehicle function evaluation and data acquisition and analysis, providing a complete set of professional tools. With reference to existing regulatory standards (e.g. European, American and national standards), corporate standards and rich experience parameters, it can support more than 400 evaluation reports.

  • Support ACC /AEB/FCW Function Evaluation.

  • Support for LCA/LDW/LKA Function Evaluation.

  • Support for active safety functions such as BSD/AP and autonomous driving functions such as AVP, high-speed autopilot and urban autopilot,and support for customized Function Evaluation.

  • Provide rich interfaces in Python, C#, etc. to facilitate users to add custom test reports.


开放道路13.jpg



7、Intelligent Connected Vehicle Function Test

ASEva Ground Truth (V2X) System realizes true value perception of vehicle-vehicle, vehicle-road and human-vehicle through advanced MESH networking and high- precision positioning technology, which is widely used in closed field and open road for system function testing.

The ASEva Ground Truth (AVP) System enables indoor and outdoor automated parking (APA) and valet parking (AVP) testing through advanced indoor positioning technology and automatically generates parking assessment reports.


开放道路14.jpg


8、Data re-injection

The ASEva data injection system injects the valid data back to the DUT after the conversion of the collected raw video and bus data in a specific format and data co-frequencying, and is used for the development and optimization of the DUT perception algorithm. The sensor benchmarking module uses the high-precision sensing results of the true value sensor to analyze the sensing results of the measured parts from multiple dimensions, find the reasons for the missed detection, false detection and wrong score of the measured parts, and provide the improvement direction for the optimization of the measured parts. The module also supports the second injection of perception results into the measurement after the data is injected back, and can compare the detection results before and after the algorithm optimization of the tested parts.

  • Support for radar CAN and RAW data re-injection.

  • Support for smart camera CAN data, video raw data re-injection and ultrasonic data re-injection.

  • Support the secondary import of data optimization results into the aseva system to do comparative analysis before and after optimization.

  • Support for customized development of SIL, HIL, VIL, DIL data re-injection.



开放道路16.jpg