Bayesian automatic model order selection in wireless communications

In wireless communications, there are many estimation problems require the information of model order. For example, the number of propagation paths in channel estimation, the number of active users in grant-free access, and the number of interference sources in radio map reconstruction. While it is common to assume the model orders are known in wireless communication research, in practice, this information is hardly known precisely. This talk will introduce how Bayesian method could solve this problem without exhaustive tuning.