We considered an energy harvesting (EH) node that periodically takes a measurement and conveys it to a destination over multiple EH relays operating in a decode-and-forward fashion, using the automatic repeat request (ARQ) protocol. Packets that are not delivered to the destination before the next measurement is taken are dropped. The objective was to design a retransmission-index based power control policy (RIP) for each node, which minimizes the packet drop probability (PDP). To this end, we first derived expressions for the PDP, in terms of the RIPs at the nodes. Next, we obtained near-optimal power control policies. The key contribution of this work shows that the design of EH links can be significantly simplified by replacing the energy neutrality constraint with a simpler average power constraint, without compromising the optimality of the solution.
Our contributions in this direction are the following:
The objective in this work was to design a power control policy that maximizes the long-term time-averaged throughput of a Gaussian multiple access channel, where the transmitting nodes as well as the access point (AP) are the energy harvesting nodes. In addition, the policy is required to facilitate the uncoordinated operation of the network, i.e., in each slot, based on their own energy availability, each node and the AP need to independently decide the amount of energy to be used. We analytically showed that time-sharing based online policies, which require no coordination among the nodes and uses time-dilation, achieve a throughput that converges the optimal throughput, asymptotically in the battery size. We also showed that a policy with occasional one-bit feedback from access point about its battery state requires a smaller battery size to achieve the same performance as a policy which operate without feedback.
Our contributions in this area are the following: