Page 55 - ATZ WORLDWIDE
P. 55

tance functions [1, 2]. This is imperative   applying the approach is on calibrating   this information to the driver assistance
            in particular for validating complex sce-  and testing functions regardless of the   function being tested. This evaluates the
            narios effectively, not least for safety and   sensors that are used.  identified road situation and, if neces-
            cost reasons as well as from the aspect                             sary, triggers an action. In the visualis-
            of reproducibility (above all highly   FUNCTIONALITY AND COMPONENTS  ation, the test engineers in the vehicle
            dynamic and complex scenarios). Fur-                                see the real-world test environment, the
            thermore, the number of test scenarios   The ViL test method is based on combin-  virtually generated target objects that
            needed for validation on the test track   ing a real-world test vehicle with virtual   move about on the real-world test track,
            and the number of real-world test kilo-  objects and scenarios. To generate a criti-  and experience the test vehicle’s real-
            metres in the field are growing as the   cal situation, virtual target objects are   world response (for example in the form
            degree of automation increases, FIGURE 1.  added to the real-world test environment.   of a visual-acoustic warning, an emer-
              The gap emerging in the validation   These are shown on a mobile device   gency braking, avoidance or stopping
            process can be bridged by means of the   (for example a tablet) in the wind-  manoeuvre). In technical terms, this is
            Vehicle-in-the-Loop (ViL) approach   shield using an augmented reality app,   realised using five components which,
            described in this paper as a way of com-    FIGURE 2. Virtual sensors detect the   apart from the central control compo-
            bining simulation and real-world tests in   superimposed objects and communicate   nent, can be exchanged or adapted to
            an appropriate manner, FIGURE 1. This
            new approach also demonstrates a way
            of permitting the progressive develop-
            ment of functions through testing and     Until now
            validation.                               Future
                                                      Future with ViL

            DEMANDS ON THE VEHICLE-IN-
            THE-LOOP APPROACH
            The basic idea in testing driver assistance   Effort
            systems in the real-world vehicle is to gen-
            erate a critical driving situation with suit-
            able tools and then evaluate the response
                                                                                                       Vehicle
            from the assistance system. To make the                                              (proving ground,
            tests reproducible and as realistic as possi-  MiL/SiL  HiL         ViL                    street)
            ble, for example in testing emergency                               Reality
            braking systems, collision-tolerant target   FIGURE 1 The effort of validating future driver assistance functions is set to increase, revealing a gap
            objects are used for representing vehicles,   between simulation-based test methods and real-world test drives which can be bridged by the Vehicle-
            pedestrians and cyclists [3].     in-the-Loop (ViL) approach (green line) in the future (© IAV)
              Increasingly complex traffic situations
            involving various and several road users
            as well as driving situations of a more
            dynamic nature are demanding new
            approaches for safe and resource-friendly
            testing. The ViL approach presented in
            this paper aims at bridging the gap
            between real-world vehicle tests and vir-
            tual scenarios and providing a way of
            seamlessly integrating this into the exist-
            ing processes for testing and validating
            driver assistance functions. In this
              context, the approach must meet the
            requirement of moving the real-world
            vehicle in a real-world test environment
            into which virtual objects are fed. The
            aim is not to embed the vehicle in a
            purely virtual environment but to ensure
            real-world conditions for the greatest
            possible number of testing components,
            FIGURE 2.
              A further demand on realising the test
            method is a modular structure that pro-
            vides maximum flexibility to allow for   FIGURE 2 The basic idea behind the ViL approach is to add virtual objects and scenarios to the real-world
            the varying conditions. The focus of   test environment (© IAV)
            ATZ worldwide  09|2017                                                                           53
   50   51   52   53   54   55   56   57   58   59   60