Abstract
This paper introduces JUDRS (Joint Uplink and Downlink Relay Selection), a scheme that selects a single optimal relay to handle both uplink and downlink communications simultaneously. By accounting for asymmetric traffic loads and weighting energy costs across different devices, JUDRS significantly improves energy efficiency and reduces signaling overhead while maintaining full spatial diversity.
Introduction
In cooperative cellular networks, mobile terminals act as relays to help transmit signals between a Base Station (BS) and a Mobile Station (MS). While previous research has extensively covered relay selection, most of it focuses on unidirectional transmission (either uplink or downlink). In reality, cellular traffic is bidirectional and often asymmetric—for example, a user might download a large file (heavy downlink) while sending only small acknowledgments (light uplink).
Existing selection methods that treat these directions independently incur high signaling overhead and fail to optimize for the total energy consumed across the entire "two-way" conversation. Furthermore, different nodes have different power constraints; a battery-powered smartphone (MS) has a much higher "cost" for energy consumption than a plug-in Base Station.
Methodology
The authors propose the Joint Uplink and Downlink Relay Selection (JUDRS) protocol. The core innovation lies in its selection criteria and its signaling process. The transmission is organized into two phases for both uplink and downlink, as shown in Figure 1.
The selection process follows a five-step sequence:
- Channel Estimation: The MS broadcasts a Request-to-Send (RTS) packet for nearby relays to estimate the MS-to-relay channel quality.
- Feedback: Relays meeting a minimum quality threshold send their Channel Quality Indicator (CQI) to the BS.
- BS Estimation: The BS estimates the relay-to-BS and direct MS-to-BS channels.
- Optimal Selection: The BS selects the relay that minimizes a weighted total energy consumption function ($E_{coop}$), incorporating traffic asymmetry ($\zeta$), device weights ($\omega$), and channel reciprocity.
- Execution: The BS broadcasts the selected relay index to begin data transmission.
The scheme was evaluated using numerical simulations with standard cellular parameters (2.5 GHz carrier, 180 kHz bandwidth, 3.76 path loss exponent) and compared against "best worse channel" and "best harmonic mean" benchmarks.
Results and Discussion
The simulations demonstrate that JUDRS consistently outperforms benchmark selection methods in terms of total energy per bit. By picking one relay for both directions, the system drastically reduces the coordination overhead.
As illustrated in Figure 5, the energy-saving gains of JUDRS become more pronounced as traffic becomes more asymmetric (i.e., as $\zeta$ moves away from 0.5). This "global" view of the traffic session leads to better overall efficiency because the system can favor a relay with a superior channel in the dominant traffic direction. Additionally, as the number of available relays increases, the energy required to transmit a single bit decreases, confirming the selection gain of the proposed algorithm. The mathematical proof also shows that JUDRS achieves a diversity gain of $N+1$, ensuring high reliability.
Conclusion
JUDRS offers a significant improvement in energy efficiency for cooperative networks by jointly optimizing for bidirectional traffic. It effectively balances the energy costs of different network nodes and adapts to varying traffic loads. Future work could explore relaxing the channel reciprocity assumption for FDD systems and investigating the impact of high-speed mobility on selection performance.
Summary updated