Information waveform manifold based preprocessing for nonlinear multichannel modulation in MIMO channel

The nonlinear multichannel modulation (particularly the CPM class) can better respect technological constraints on the transmitter (e.g. nonlinearity) due to its multidimensional waveform space which allows to decouple the second order properties of the constellation and the actual waveform shape (e.g. constant envelope). However the receiver processing complexity, especially in MIMO channel, is very high due to this waveform multidimensionality. We present an information waveform manifold (IWM) based preprocessing reducing the dimensionality of the problem. It is based on nonlinear projector, optional processing on IWM and decoding isomorphism with channel symbols expansion space. We show that the nonlinear projector is generally suboptimal, however the information loss comparing with optimal processing is negligible. The decoding isomorphism also directly separates the individual components of the received signal in MIMO channel