Determining user specific spectrum usage via sparse channel characteristics

We consider a device-to-device scenario, where multiple users transmit complex symbols with constant amplitude on exclusive OFDM resources. However, the spectrum usage and therefore the assignment of resources by each user is unknown at the receiver. The number of resources allocated by each user are well chosen, so that the sparse channel characteristics can be coarsely estimated once the correct resource assignment is known. The receiver faces an inverse problem with two unknowns, the resource allocation and the spectral channel power. We propose an objective function, which identifies the correct resource assignment by just observing the received signal and exploiting the sparse channel properties. We show by simulations based on a practical-relevant ITU channel model, that the objective function is well chosen and one can recover the user allocations under the presence of noise. These results enable us to identify non-adjacent spectrum usage belonging to the same transmitter which goes beyond edge-detection strategy usually used in this context.