In the real world, a dynamic and unpredictable environment, an “Advanced Driver-Assistance System” (ADAS) should be safe for itself, pedestrians, and other obstacles. One possible approach to ensure the safety of ADAS is “Runtime Monitoring” (RM). In this context, additional formal safety mechanisms are added to the system. This thesis aims to explore RM for ADAS. Our main contributions consist of three parts. Firstly, we conducted a rapid review to find relevant RM techniques for ADAS. We provided detailed information on our search query and the filter criteria for the papers. We extracted the information from 16 remaining relevant papers and applied an existing taxonomy to classify the techniques. Secondly, we defined the hardware criteria given by the “Robot Operating System” (ROS)-based “Autonomous Research Vehicle” (ARV) called Mecabot TX, the use case for the prototype, and four safety requirements for our runtime monitor. In our use case, the system performed “Advanced Emergency Braking” (AEB) without the intervention of a driver. Based on that, we then chose one of the 16 techniques for our prototype. The technique we evaluated to be optimal was rtamt that relied on specifications written in “Signal Temporal Logic” (STL). Thirdly, we implemented one passive runtime monitor for each safety requirement using rtamt. We proposed an architecture consisting of an object tracker, the AEB logic, and the runtime monitors. We verified our monitors and conducted an experiment to test the implementation. We identified that the runtime monitors successfully detected multiple violations for each safety requirement during the test run. By analyzing the violations, we gained helpful insights for debugging the system and improvements for its safety. Therefore, our work paves the way for future research in the area of RM for ADAS.