TL;DR
This paper tackles the challenge of maximizing data throughput in wireless relay networks powered by energy harvesting. It introduces a novel concept of "energy diversity" that can be exploited by delaying data, leading to significant throughput gains compared to systems with strict delay constraints, even over static channels.
What Problem Does This Paper Address?
Wireless communication systems, especially sensor networks, are increasingly relying on energy harvesting (EH) to power devices, offering a sustainable alternative to finite batteries. However, EH sources are intermittent, meaning the energy available for transmission fluctuates over time. This creates a challenge for optimizing communication strategies, as devices must operate within their harvested energy budgets. This paper focuses on a three-node Gaussian relay channel where both the source and the relay are powered by EH. The core problem is to maximize the total data throughput over a finite time horizon, considering two distinct traffic types: delay-constrained (DC) and no-delay-constrained (NDC). The DC traffic requires immediate decoding, while NDC traffic allows for flexible delays, opening up possibilities for exploiting the intermittent nature of EH more effectively.
Methods
The authors propose novel algorithms for optimizing power allocation in both DC and NDC scenarios.
- Delay-Constrained (DC) Case: They formulate a convex optimization problem and develop an efficient algorithm based on KKT conditions and the observation that optimal power allocation is non-decreasing over time. This algorithm is a forward search in time, considering both source and relay energy profiles.
- No-Delay-Constrained (NDC) Case: While the problem is non-convex, they prove the optimality of a separation principle. This allows them to decouple the problem into two subproblems: optimizing source power allocation first, and then optimizing relay power allocation. This leads to a two-stage power allocation algorithm.
- Energy Diversity: A key conceptual contribution is the identification and exploitation of "energy diversity." This phenomenon arises from the independent energy availability at the source and relay over time, allowing for higher throughput in the NDC case by strategically delaying data to better utilize harvested energy.
Datasets
Not discussed in the paper. The paper focuses on theoretical analysis and algorithm development, with numerical results generated based on assumed energy profiles.
Key Insights and Ideas
- Energy Diversity is Real: The paper demonstrates that even with time-invariant channels, independent energy harvesting at the source and relay can create a form of "energy diversity." This diversity can be exploited to improve throughput, especially when delay is not a critical constraint.
- Delay Tolerance Pays Off: Allowing for flexible decoding delays (NDC) significantly enhances system performance compared to strict delay constraints (DC). This is because it enables better utilization of harvested energy, smoothing out the impact of intermittent power availability.
- Separation Principle for NDC: The authors prove that for the NDC case, the problem of optimizing source and relay power allocation can be separated. This simplifies the optimization process considerably, making it more tractable.
- Monotonic Power Allocation: A crucial finding is that optimal power allocation strategies for both DC and NDC cases tend to be non-decreasing over time. This property is leveraged in the development of efficient algorithms.
Key Results
- DC Case: An efficient algorithm is developed to compute optimal joint source and relay power allocation for delay-constrained traffic.
- NDC Case: A two-stage algorithm is proposed based on a proven separation principle, enabling efficient computation of optimal power profiles for no-delay-constrained traffic.
- Throughput Gains: Numerical results show that the NDC scheme can achieve significantly higher throughput than the DC scheme, especially when the direct link gain ($h_0$) is small. This gain is attributed to the exploitation of energy diversity.
- Throughput Limit: For the NDC case, a throughput limit is achieved even at very small direct link gains, whereas the DC case shows a more linear increase with the direct link gain.
Limitations and Future Work
- Deterministic EH Model: The paper assumes a deterministic EH model, where energy arrival times and amounts are known in advance. The authors suggest that extending the analysis to random EH models (e.g., block-Markov models) would be a valuable future direction for more practical scenarios.
- Infinite Energy Storage: The analysis assumes infinite energy storage capacity. Considering finite energy storage, as explored in prior work, is identified as an important extension.
- DF Relaying: The study focuses on the Decode-and-Forward (DF) relaying scheme. Future work could explore other relaying techniques like Amplify-and-Forward (AF) and Compress-and-Forward (CF), as well as more general relay channel models.
Summary updated