Researchers at the Institute of Control Engineering investigate methods and concepts for the safety assurance of autonomous road vehicles. Simulation is a key tool here, as it allows vehicles to be tested in many traffic scenarios safely and at low cost. Simulation results, however, are only meaningful if the simulation is valid for the scenario being tested: the models used must cover the conditions that occur in it, such as speeds, road conditions or weather.
The institute describes these relationships with contracts: the validity of each simulation model, and the requirements of each test case, are stated as conditions that can be checked automatically. So far, however, test case contracts are written by hand. At the same time, test cases are usually derived from scenarios that are already described systematically, for example in ontologies or in the ASAM OpenSCENARIO format. In widely used simulators such as CARLA, it is also hardly checked whether a simulation run actually reproduced the intended scenario and stayed within the valid range of the built-in models.
In this master’s thesis, test case contracts shall be derived automatically from scenario descriptions and checked in CARLA. First, a literature review on scenario ontologies, scenario description formats and contract-based design is to be conducted. Next, a mapping shall be developed that translates scenario concepts into the variables used in contracts. Based on this, test case contracts are to be derived from logical scenarios and test objectives. The scenarios shall then be executed in CARLA via OpenSCENARIO, recorded, and the contracts monitored on the recorded runs (programming in Python). A proof of concept with at least two scenarios shall show how well the derived contracts match expert-written ones, followed by a critical evaluation. The thesis is part of a work package with a second master’s thesis; both can be worked on independently, and collaboration is welcome.
Required are...
Helpful, but not strictly required are...