Source-linked AI summary
Spectrum Sharing Scheme Between Cellular Users and Ad-hoc Device-to-Device Users
Brett Kaufman, Jorma Lilleberg, Behnaam Aazhang
TL;DR
The paper asks how ad-hoc D2D users can share fully utilized licensed cellular spectrum without violating cellular interference constraints. It develops distributed statistical power control and random-access route discovery, finding lower route-discovery failure and transmission overhead, low D2D outage, small cellular-performance loss, and significant D2D power savings.
Problem
Spectrum scarcity and underutilization motivate sharing licensed cellular spectrum, but interference must remain within limits while D2D communication remains reliable.
Method
The paper develops a distributed protocol that estimates channel gains for D2D power control and uses network-informed random-access discovery to establish single-hop or multi-hop routes.
Results
Network information lowers route-discovery failure and transmission requirements; the protocol achieves low D2D outage, significant D2D power savings, and large D2D improvements with only a small macro-user performance loss.
Takeaways & Limitations
D2D users can simultaneously communicate over the same frequency resources as a fully loaded cellular network using distributed power control and route discovery without base-station coordination.
Abstract
from arXiv · showhide
In an attempt to utilize spectrum resources more efficiently, protocols sharing licensed spectrum with unlicensed users are receiving increased attention. From the perspective of cellular networks, spectrum underutilization makes spatial reuse a feasible complement to existing standards. Interference management is a major component in designing these schemes as it is critical that licensed users maintain their expected quality of service. We develop a distributed dynamic spectrum protocol in which ad-hoc device-to-device users opportunistically access the spectrum actively in use by cellular users. First, channel gain estimates are used to set feasible transmit powers for device-to-device users that keeps the interference they cause within the allowed interference temperature. Then network information is distributed by route discovery packets in a random access manner to help establish either a single-hop or multi-hop route between two device-to-device users. We show that network information in the discovery packet can decrease the failure rate of the route discovery and reduce the number of necessary transmissions to find a route. Using the found route, we show that two device-to-device users can communicate with a low probability of outage while only minimally affecting the cellular network, and can achieve significant power savings when communicating directly with each other instead of utilizing the cellular base station.
I. INTRODUCTION
The paper addresses spectrum scarcity by enabling distributed D2D users to opportunistically share fully utilized cellular spectrum while limiting interference to licensed links. It combines statistical power control with route discovery to support single-hop and multi-hop communication.
- Spectrum scarcity and the high cost of new spectrum motivate dynamic sharing and spatial reuse techniques.
- The central challenge is keeping interference received by cellular users within the allowed interference temperature while maintaining reliable D2D performance.
- Statistical channel-gain estimates set D2D transmit powers, after which random-access discovery establishes routes between D2D users.
- Network information in discovery packets lowers route-discovery failure and reduces the transmissions needed to find a route.
- The system models seven circular cells with NC orthogonal channels, one active macro user per channel, and no spare channels for redistribution.
- D2D users form a distributed network within a small cluster, communicating directly over one or more hops without base-station assistance.
B. Channel Model
The channel model represents links with distance-dependent path loss, Rayleigh fading, and additive white Gaussian noise, while distinguishing macro, source, and destination parameters.
- The model uses Euclidean distance and pathloss exponent α for large-scale fading, Rayleigh fading for small-scale effects, and additive white Gaussian noise.
- Channel gain is defined as h_ij = |f_ij|^2, with transmit powers and pathloss terms specified for desired and interfering links.
- All users share noise power σ^2, and users are assumed to know the pathloss of their respective base-station links for interference management.
III. DEVICE-TO-DEVICE COMMUNICATION
The proposed protocol lets D2D users share frequency resources with active uplinks when their interference remains within the cellular network’s permitted margin. Its two stages are power control followed by route discovery.
- D2D links opportunistically reuse cellular frequency resources only when they do not reduce the cellular SINR beyond the allowed margin.
- The protocol first calculates an allowed D2D transmit power and then discovers a single-hop or multi-hop route to the destination.
A. Power Control for Device-to-Device Users
D2D power control uses estimates of interference-channel and desired-link gains to constrain cellular interference while improving D2D link quality. The paper also analyzes idealized inversion cases to bound performance.
- The D2D transmit-power bound is derived from the macro link’s required SINR and the interference caused at the base station.
- Statistical estimation of the source-to-base-station gain enables decentralized power selection without coordination from the base station.
- Underestimating the source-to-base-station gain makes the calculated maximum D2D power exceed the true maximum, causing macro-user outage.
- Overestimating the source-to-base-station gain yields a D2D power below the true maximum, conservatively reducing possible transmission power.
- The desired D2D link can receive additional power control based on an estimate of h_SD, with channel inversion used to improve received power.
- Using d̂h_SD = E[h_SD] provides a practical low-overhead estimate that controls pathloss while ignoring fast-fading effects.
- Perfect and truncated channel inversion are analyzed as impractical reference cases for bounding system performance.
- The same controlled power is applied hop by hop by substituting the transmitting and receiving D2D nodes for the source and destination.
B. Distributed Route Discovery for Two-way Device-to-Device Communication
The protocol discovers two-way single- or multi-hop D2D routes by exchanging transmission and interference information, allowing nodes to screen links before forwarding. This information improves route discoverability and reduces unnecessary transmissions.
- Protocol design: DSR floods source-initiated discovery packets while CSMA/CA allows only one D2D user to access the channel at a time.Discovery packets exchange relay addresses, and flooding rules prevent repeated traversal of the same link.
- Protocol design: Forwarding relays append their transmission power and measured interference power to discovery packets alongside their addresses.The source begins the process with its calculated transmit power and measured interference power.
- Link screening: Each receiving node uses packet information and measured SINR, interference, and noise to determine whether a two-way link with the source is feasible.The node estimates the combined pathloss and channel gain, then evaluates the required return-link transmit power under the D2D SINR constraint.
- Route construction: Nodes forward discovery packets only when they know a two-way link exists, and the destination collects relay nodes forming a multi-hop route.A node can serve as a relay when a single-hop route does not exist; otherwise, it can be the destination of a direct route.
- Route construction: Strong macro-user interference can prevent a two-way route through a relay, whereas sufficiently favorable interference conditions permit a single-hop route.The example shows D1 receiving directly while a route involving R1 is rejected because the macro user causes excessive interference.
- Protocol benefits: Including transmission and interference powers in discovery packets improves the likelihood of discovering a two-way route and lowers the required number of transmissions.The source can therefore determine that a route exists to the intended destination before data delivery.
C. Route Discovery Simulation Results
Simulations evaluate route discovery over random network topologies and show that network information improves two-way route discovery while reducing discovery overhead. The benefits vary with pathloss and relay population.
- Failure probability: Network information in discovery packets significantly improves the probability of discovering a two-way route.The comparison is made between discovery with and without network information.
- Failure probability: As the pathloss exponent α increases, the probability of route-discovery failure P_fail^D decreases.Higher pathloss also reduces interference from cellular users, allowing D2D users to transmit at higher power despite greater per-hop attenuation.
- Transmission savings: T_save shows increasing transmission savings as α decreases when network information is included in discovery packets.T_save compares average transmissions with network information, T_NI, against the average without it, T.
- Transmission savings: For small α and larger N_D, network information yields more savings because interference creates more one-way links and prevents relays from forwarding packets.The effect is strongest when many relays might otherwise forward discovery packets received over one-way links.
IV. OUTAGE ANALYSIS OF DEVICE-TO-DEVICE COMMUNICATION
The paper derives D2D outage probability geometrically from random user distances and fading, then distinguishes tractable analysis from simulations that include neighboring-cell interference.
- Analytical model: The outage analysis combines distance and fading distributions to account for random D2D locations and channel conditions.The resulting expression gives outage probability for a link between randomly placed D2D users.
- Analytical model: The single-hop analytical expression ignores interference from surrounding cells because incorporating neighboring-cell distances makes the solution intractable.Those neighboring-cell interference effects are included in the simulation results.
A. Single-hop Probability of Outage Derivation
The paper derives single-hop D2D outage by converting the SINR distance constraint into a geometric coverage probability, then averaging over channel gains and network distances. Because the resulting expression depends on nine random variables, the final analytical results are numerically approximated and verified through simulation.
- Channel averaging: For multiple orthogonal channels, outage on each channel is modeled as independent and identically distributed across channels.The bandwidth is divided into NC orthogonal channels, and D2D users may access any channel.
- Outage geometry: The single-hop outage analysis uses a randomly placed D2D source-destination pair and defines outage through the maximum feasible separation dmax.A link succeeds when the destination lies within the source’s coverage region and the D2D SINR exceeds βD.
- Outage geometry: The intersection-area expression handles both intersecting circles and cases where one circle is completely contained within the other.Although the formula may return a complex value under the intersection assumption, its real component equals the physically relevant overlap area.
- Outage geometry: The success probability Pr[dSD ≤ dmax] is obtained as the ratio of feasible destination locations to the cluster area πr^2.The feasible region is the intersection between the source’s coverage region and the D2D cluster.
- Statistical averaging: The joint distribution separates channel gains from distances because Rayleigh fading terms are independent of the network distances.The six-variable distribution is decomposed into a channel-gain factor and a distance factor, with channel gains modeled as independent and identically distributed exponential variables.
- Statistical averaging: The final outage probability averages over nine random variables, but its closed-form evaluation is intractable and therefore requires numerical approximation and simulation verification.The distance distributions account for the randomly located cluster, D2D source, macro user, and their geometric dependencies within the macrocell.
B. Bounds on the Single-hop Probability of Outage
The paper evaluates single-hop D2D outage under varying channel diversity, cluster size, and pathloss, comparing perfect and statistical channel-gain estimates. More channels and higher pathloss generally reduce D2D outage, while the cellular outage remains mostly below 10^-2 under the considered settings.
- Single-hop D2D outage: Analytical and simulated outage results are compared for a single D2D link sharing N_C channels with active macro users.The evaluation uses r/R = 0.25 and considers perfect and statistical channel-gain estimates.
- Single-hop D2D outage: As N_C increases, P_out^D decreases because D2D users gain more channel diversity and can select channels with lower macro-user interference.Analytical results match the simulations closely.
- Single-hop D2D outage: As r/R decreases, D2D outage decreases because the maximum source-destination distance becomes smaller.Fading becomes more dominant than pathloss as the endpoints move closer, increasing the gap from the statistical estimate to the lower bound.
- Single-hop D2D outage: As pathloss exponent α increases, D2D outage decreases because D2D users can transmit at higher power and receive less macro-user interference.The higher pathloss also reduces D2D interference at the base station.
- Macro-user outage: Most macro-user outage values are below 10^-2 and level off for large α.The study varies N_C and r/R while comparing perfect and statistical D2D power-control estimates.
- Outage tradeoff: D2D outage improves by about a factor of 100, while macro-user loss remains below a factor of 10 in the compared outage tradeoff.Higher D2D outage from less accurate channel estimates corresponds to lower macro-user outage.
C. Multi-hop Simulation Results
The multi-hop evaluation compares route power with cellular-mode power and examines D2D and macro-user outage. Multi-hop D2D communication provides substantial savings, especially in high-pathloss environments.
- Power comparison: Cellular-mode power sums transmission from the D2D source to the base station and from the base station to the destination.D2D-mode power sums the transmit powers across the N_Hops route.
- Power savings: High pathloss environments show significant savings for multi-hop D2D routes compared with cellular mode.Savings are meager in low-pathloss environments because the D2D user may have a strong channel to the base station.
PTSB + PTBD
The route-length analysis studies how the number of D2D relays affects the average hops between random source-destination pairs. Routes generally remain short at higher pathloss, whereas high interference can require longer routes.
- Route length: N_D = 0 denotes a single-hop route between the source and destination.The route analysis varies the number of D2D relays in the cluster.
- Route length: For α ≥ 3, route length quickly saturates at two hops or fewer even as N_D becomes large.Usually one or two relays suffice to establish a two-way D2D route.
- Route length: At α = 2, routes often span more hops in the high-interference scenario.The source-destination distances are generally not much larger than the single-hop distance.
V. CONCLUSIONS
The paper presents a distributed protocol for D2D communication over cellular spectrum, combining interference-aware power control with route discovery. It reports improved discovery efficiency, near-bound outage performance, significant power savings, and identifies coordination as an extension.
- The protocol enables D2D users to share frequency resources with a fully loaded cellular network without base-station coordination.
- Power control first limits D2D interference, after which discovery packets establish routes to intended destinations.
- Network information in discovery packets lowers route-discovery failure probability and reduces the transmissions needed to find a destination.
- Given a route, practical statistical power control performs near the lower bound derived using perfect channel inversion, while D2D gains incur only a small macro-user performance loss.
- Simulation results show significant power savings for D2D routes compared with communication through the cellular base station.
- More base-station coordination and dynamic channel estimation are proposed to support simultaneous D2D clusters and improve interference control.