Joint Channel Estimation and Data Detection for ODMA in Unsourced Random Access

For the massive machine-type communication （mMTC） scenario in the Internet of Things （IoT）， unsourced random access （URA） is an effective potential technology that supports uncoordinated access for a large number of users based on a shared codebook mechanism. In existing URA decoding schemes， channel estimation and user data detection are mostly implemented separately. The error propagation of estimation and detection affects the information detection and decoding performance of the system in the presence of high load or low signal-to-noise ratio. To solve this problem， herein we proposed a joint channel estimation and data detection algorithm based on on-off division multiple access （ODMA） in the URA system. This algorithm adopted bilinear generalized approximate message passing （BiG-AMP）， which enabled the joint estimation and detection of channel parameters， active patterns， and user data by alternately updating between the channel and signal matrices. In addition， an initialization strategy assisted by a “pilot” was introduced to enhance the convergence speed and numerical stability of the algorithm. Numerical simulation results showed that within the given range of the number of active users and the pilot length， the normalized mean square error （NMSE） of the channel estimation and data detection of the proposed algorithm was superior to that of separated channel estimation and data detection algorithms. Compared with the classic separated URA scheme based on the pilot， this algorithm afforded a performance gain of up to 1.1 dB.