Abstract : Recently, several studies have been moving towards the idea to adopt massive arrays for localization and mapping applications thanks to the high scanning resolution achievable. In this framework, this paper investigates the impact of different design parameters in order to maximize the mapping performance while keeping the receiver architecture as simple as possible when a Bayesian approach is used. To this end, thresholding strategies to filter out unwanted received signal components (cleaning operation before soft mapping), e.g. thermal noise, are analyzed together with other design parameters that can operate directly on raw data. Considerations about possible system design configurations are then drawn inspired by results obtained using measured data, which showed a more pronounced impact of the cleaning approach for 1 GHz bandwidth with respect to 3 GHz.
