obtaining first sensor outputs that represent actual states of a dynamic object using a first sensor that is located in the dynamic object;

storing actual state values for the dynamic object that are based on the first sensor outputs;

obtaining a second sensor output using a second sensor that is located in a subject vehicle, wherein the second sensor output represents an observation of the dynamic object;

determining estimated locations of the dynamic object based on the second sensor output using a sensing algorithm;

storing estimated location values corresponding to the estimated locations of the dynamic object;

defining training samples that include the actual state values, the actual state values including an actual location value for the dynamic object at a time point, a respective one of the estimated location values corresponding to the time point, and an actual error value that is defined as a difference between the actual location value and the respective one of the estimated location values;

training a machine learning model using the training samples such that the machine learning model is configured to determine a condition-dependent error distribution that is conditioned on inputs that represent simulated states of a simulation and represents an error during determination of estimated locations of simulated dynamic objects; and

subsequent to training the machine learning model, performing one or more iterations of a simulation that includes: