Many advanced driver assistance systems (ADAS) are already firmly anchored and established in everyday life. Emergency braking systems, adaptive cruise control or lane departure warning and lane change assistants ensure greater safety on the road. Gradually, more and more complex assistance functions are being added, which are steadily making autonomous driving (AD) possible.
But how can these highly complex, automated and autonomous functions be resiliently tested? And how do you do justice to the variety of situations in the real world? We asked our egineering projects to find out more.
ADAS/AD testing
The future belongs to automated and autonomous driving functions – and our solutions are used to test them.
Scaling required?
Fast feedback via scalable test execution
Agile working methods, faster product cycles and increasing system complexity create new challenges for testing. Results must be available quickly and in large numbers. However, this often exceeds the capacity of conventional infrastructures. Our solutions therefore scale with the test volume – from individual systems to distributed test systems in clouds and clusters.
Only with a comprehensive overview of the current product development status can well-founded decisions be made. Manual system and acceptance tests in the vehicle and on test benches help here, but are time-consuming and available only to a limited extent. We therefore complement these approaches with virtual testing in SiL and MiL environments.
Our tool chain enables both single user systems and parallel execution on distributed systems with any number of physical machines, virtual machines or containers. We rely on technologies such as Docker and Kubernetes to support both internal clusters and public clouds. We scale dynamically depending on the desired test throughput.
And to avoid losing track of the flood of data, all results are compiled and analyzed centrally in test.guide – so that you can use test resources where they are needed. Of course, we will not leave you on your own in this process. Our engineering teams support you in all tasks, from initial cluster setup to operation in the cloud and data analysis.
2020-08-25
ADAS/AD testing in virtual test environments Jan Richter, tracetronic Inc.
Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD) are often seen as key factors when we talk about traffic safety. But how can you be sure whether exactly these features actually increase security? An important point here is extensive testing during the whole development process.
Scenarios preferred?
Realistic testing with scenarios
ADAS/AD functions must work reliably under a wide range of traffic and environmental conditions. Failure can lead to extremely dangerous situations. The validation of such functions therefore requires broad-based and realistic simulations of the environment and traffic. We support you with our scenario-based workflow, from the provision of the scenarios to functional release.
Our workflow starts with the selection of suitable scenarios from a database. This is used for central storage, efficient searches and the cross-project exchange of scenario data. We support both open and proprietary formats. The selected scenarios are now integrated into test cases and automatically executed and analyzed using our tools ecu.test und test.guide.
Flexibility is very important to us. Created scenarios can be combined with any test environment and analysis. This reduces development time and means that the tests can grow with the driving function – from the first model to integration in the vehicle. Since the function to be tested must be able to cope with every conceivable situation in reality, the tests should also cover a high number of variants. Our solutions provide extensive possibilities for the variation of scenarios – from classic parameterization to intelligent search and optimization algorithms. Powerful analysis capabilities help you determine coverage and get to the bottom of the behavior of your function in critical situations.
We regard scenario-based testing as an important method on the path to autonomous driving. We are therefore actively promoting these approaches in research and standardization projects. Our products support industry standards such as OpenSCENARIO and OpenDRIVE.
2020-09-30
What to expect when scenario-based testing? Jan Richter, tracetronic Inc.
With the increasing complexity of automated and autonomous driving functions, the variety of the required test scenarios and the need to automate the tests increases. Our colleague Jan Richter has dealt with this topic and writes about the topic "scenario-based testing".
Video favored?
Scenario-based testing with CarMaker
In the video on the right, our colleague Matthias explains how to automate scenario-based tests with ecu.test.
As an example, he uses IPG CarMaker and a logical scenario in which the emergency brake assistant of an ego vehicle is to be tested in detail when changing lanes on the highway.
Velocity and distance are the key parameters here, which can be easily varied using ecu.test.
Curious to see how it works? Take a look.