Errors in cyber-physical systems often have fatal consequences. One effective approach to ensure that these systems behave correctly in all cases is runtime monitoring. With runtime monitoring, a monitor continuously checks a specification against the behavior of a system during operation.

However, a problem that most monitoring approaches overlook is that the bandwidth between the system and the monitor is in reality not infinite. The bandwidth is limited because the system has restrictions for example concerning energy consumption or weight in an autonomous aircraft. Although bandwidth is limited, sensors of an autonomous aircraft always produce new values in the same frequency, regardless of their immediate necessity. To make the best use of the limited bandwidth, it is beneficial to prioritize and query sensors more often which are more essential at any given time.

In this thesis, we want to investigate this idea and show our approach with RTLola. We propose a modification to the RTLola monitoring language, so that input streams do not constantly receive new values from sensors, instead the monitor asks for new values when needed. To enable the monitor to request values, we allow the user to add timing information to triggers and output streams when the monitor should update them. Based on this information, we analyze the specification and calculate the appropriate query intervals for each stream at any given point in time. In the end, we want to evaluate our monitoring approach and compare it to the original version of RTLola.