Data Rate Performance of Mobile Multiuser MIMO Underwater Acoustic Communication Systems

This article presents a study of the Multiuser Multiple-Input Multiple-Output (MU-MIMO) Method for the multiple access of multiuser mobile underwater acoustic communications. Orthogonal multiple access schemes, such as time-division multiple access (TDMA), frequency-division multiple access, or code-division multiple access, suffer from limited data rate due to an orthogonal sharing of communication resources. In an MU-MIMO system, each user is viewed by the receiver as an input of a conventional multiple-input multiple-output (MIMO) channel using spatial multiplexing. The receiver, having multiple hydrophones, can then exploit the spatial diversity to decode the input streams of each user through successive cancellation of interstream interference. However, unlike traditional MIMO decoding, in a mobile context, each user can have its own motion-induced Doppler scale factor that needs to be efficiently treated at the receive side. First, a communication system is presented in detail, where Doppler scale domain equalization is proposed to cope with multiple mobile users having each their own Doppler scale factor. Then, a data rate performance evaluation of MU-MIMO shows its advantages compared to typical orthogonal multiple access schemes. Performance evaluation is carried out for a two-user configuration in terms of the theoretical achievable data rates computed on channels sounded at-sea and by the evaluation of effective rates obtained through channel replay of the full modulation-demodulation stages on the same sounded channels. Achievable rates of MU-MIMO with successive interference cancellation are compared with single user and TDMA, set as upper and lower performance bounds, respectively. For the studied configuration, results show that a data rate gain of 64% over TDMA is expected in MU-MIMO. Finally comparisons with the effective rates are provided for both the static and mobile users. They reveal a gain of the same level as that expected in theory at signal-to-noise ratio (SNR) above 12 dB for static users. The price to pay is a robustness loss of MU-MIMO at low SNR. In the mobile scenario, results show an additional loss in robustness compared to the static MU-MIMO case, so that the equivalence with the theoretical performance is obtained at SNR above 15 dB.