Spatial-Temporal Water-Filling power allocation in MIMO systems with harvested energy

This paper presents the power allocation solution in single link Multi-Input Multi-Output (MIMO) wireless communication systems powered by harvested energy. We first derive the channel capacity with non-causal full information in a finite time horizon, which is shown to be a convex optimization problem. The optimal power allocation can be obtained through Spatial-Temporal Water-Filling (ST-WF) solution, by which harvested energy is first allocated among antennas and then among time slots. For the causal information condition, we propose three suboptimal policies, namely constant water level policy, capacity-aware water level policy and energy-aware water level policy. Their performance is evaluated by numerical simulations, which show that the energy-aware water level policy is the best among them.