Lane Change Assist Development with Simulink

IDIADA Fahrzeugtechnik GmbH, the German Business Unit of Applus+ IDIADA, is an engineering company that provides design, testing, engineering, and homologation services to the automotive industry. We are interested in the verification and validation of advanced driver assistance systems. To optimize and validate our processes and testing capabilities, we need ADAS, whose advantages and disadvantages are known to us.

Instead of starting from scratch and to save time, we used Automated Driving Toolbox™ and Simulink^®. The default examples like ACC and LKA, with the included scenario reader, sensors, multi-object tracker, and vehicle and driver models, are a good introduction to ADAS. These parts were the basis for the lane change assist. There were some challenges on the way, such as reading lanes behind the ego vehicle, sensor fusion of lanes and objects, sensor fusion of objects behind and in front of the ego vehicle, and trajectory planning and modification in real time.

As a result, we got a lane change assist, which detects faster objects on the target lane coming from behind or slower objects ahead, waits until the object has passed, and plans the trajectory for the lane change. The trajectory was realized by two clothoid curves, which are curves with linear change in curvature and used in road design. Additionally, this LCA was built modularly by using Simulink subsystems with different sensor configurations. To verify the LCA, we automatically generated code for different scenarios with different parameters, like delta speed and distance. The Driving Scenario Designer app was the basis for the automated scenario generation. We want to implement this and other ADAS on our driving simulator DiM 250 and on the IDIADA ADAS platform tool (IDAPT) in real cars, like our CAV (level 4 taxi), to test it on our proving ground.