We consider the use of energy harvesters, in place of conventional batterieswith fixed energy storage, for point-to-point wireless communications. Inaddition to the challenge of transmitting in a channel with time selectivefading, energy harvesters provide a perpetual but unreliable energy source. Inthis paper, we consider the problem of energy allocation over a finite horizon,taking into account channel conditions and energy sources that are timevarying, so as to maximize the throughput. Two types of side information (SI)on the channel conditions and harvested energy are assumed to be available:causal SI (of the past and present slots) or full SI (of the past, present andfuture slots). We obtain structural results for the optimal energy allocation,via the use of dynamic programming and convex optimization techniques. Inparticular, if unlimited energy can be stored in the battery with harvestedenergy and the full SI is available, we prove the optimality of a water-fillingenergy allocation solution where the so-called water levels follow a staircasefunction.
展开▼
