Training signal design for channel estimation in decode and forward relay networks

We provide studies on the training based channel estimation for decode-and-forward relay networks. Since multiple relay nodes are geographically distributed over the service region, channel estimation is different from the traditional way in that each relay has its own individual power constraint. Since the general optimization for the minimum mean square error based channel estimation is hard to solve, we consider three special yet reasonable scenarios. The problem in the first scenario lies in the so called semidefinite programming and could be efficiently solved by state of the art optimization tools. Closed-form waterfilling type solutions are found in the remaining two scenarios, of which the first one has a special physical meaning and its corresponding structure is named as cave-filling.