PaperlayerView exact revision on arXiv

Source-linked AI summary

Coordinated Dynamic Spectrum Management of LTE-U and Wi-Fi Networks

Shweta Sagari, Samuel Baysting, Dola Saha, Ivan Seskar, Wade Trappe, Dipankar Raychaudhuri

arXiv:1507.06881v1cs.ITcs.NIcs.PF

TL;DR

The paper examines interference between Wi-Fi and LTE sharing emerging unlicensed bands. It models and experimentally validates coexistence, then proposes logically centralized coordination using power control and time division, which improves throughput and fairness across the networks.

  • Problem

    Wi-Fi and LTE must coexist in shared unlicensed spectrum, but uncoordinated operation causes significant mutual interference and performance degradation.

  • Method

    The paper develops and partially validates an analytical interference model, then uses an SDN-enabled logically centralized framework for coordinated Wi-Fi–LTE power control and time division.

  • Results

    Joint power control and time division increase Wi-Fi and LTE throughput, with 10 percentile throughput reaching 15−20 Mbps versus approximately zero without coordination.

  • Takeaways & Limitations

    Coordinated Wi-Fi–LTE spectrum management can improve aggregate throughput while making the networks' access to shared spectrum more comparable.

Abstract

from arXiv · show

This paper investigates the co-existence of Wi-Fi and LTE in emerging unlicensed frequency bands which are intended to accommodate multiple radio access technologies. Wi-Fi and LTE are the two most prominent access technologies being deployed today, motivating further study of the inter-system interference arising in such shared spectrum scenarios as well as possible techniques for enabling improved co-existence. An analytical model for evaluating the baseline performance of co-existing Wi-Fi and LTE is developed and used to obtain baseline performance measures. The results show that both Wi-Fi and LTE networks cause significant interference to each other and that the degradation is dependent on a number of factors such as power levels and physical topology. The model-based results are partially validated via experimental evaluations using USRP based SDR platforms on the ORBIT testbed. Further, inter-network coordination with logically centralized radio resource management across Wi-Fi and LTE systems is proposed as a possible solution for improved co-existence. Numerical results are presented showing significant gains in both Wi-Fi and LTE performance with the proposed inter-network coordination approach.

I. INTRODUCTION

Growing mobile data demand is driving use of additional spectrum, small cells, and more efficient spectrum utilization, including shared unlicensed bands. Because Wi-Fi and LTE can interfere when co-channel networks overlap, the paper develops interference models and proposes SDN-enabled coordination using power control and time division.

  • Motivation: 2.4 and 5 GHz unlicensed bands, among other candidate bands, are being considered for mobile and fixed wireless broadband services and unlicensed LTE operation.The paper also discusses 3.5 GHz, 60 GHz, and TV white spaces for small-cell or backhaul uses.
  • Motivation: Shared unlicensed bands place Wi-Fi and LTE in the same frequency, time, and space, creating mutual interference and degrading system performance.The paper focuses on coordinated coexistence between these two prominent radio access technologies.
  • Contributions: An SDN-enabled coordination framework supports dynamic spectrum management across multi-operator, multi-technology networks without requiring changes to existing radio standards or protocols.The architecture provides broader visibility into technologies, spectrum bands, clients, and operators for coordination.
  • Contributions: The paper introduces an analytical model for Wi-Fi–LTE interference and partially validates it experimentally using USRP-based LTE nodes and COTS IEEE 802.11g devices in the ORBIT testbed.The model characterizes coexistence while networks share the medium in time, frequency, and space.
  • Contributions: The proposed optimization coordinates Wi-Fi and LTE through power control and time division channel access, targeting aggregate throughput while considering per-client throughput requirements.The paper evaluates this framework for improved coexistence between the two networks.

II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE

Wi-Fi and LTE differ in medium access and transmission behavior, making co-channel coordination difficult. The paper models their interference and throughput using SINR, Wi-Fi CCA/CSMA behavior, and LTE scheduling characteristics.

  • II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE: Different MAC designs make coordination between Wi-Fi and LTE challenging in shared spectrum.Wi-Fi uses distributed carrier sensing, while LTE schedules transmissions across time and frequency.
  • II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE: The model focuses on a single Wi-Fi and LTE cell sharing one channel and bandwidth, providing a building block for more complex deployments.The assumed downlink scenario has one client and saturated traffic at each AP without MIMO.
  • II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE: Interference is represented through SINR, where received power depends on transmit power and channel gain while the other network contributes interference and noise.Channel gain incorporates distance-dependent path loss and additional antenna, cable, and wall-loss factors.
  • II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE: Throughput is modeled as α_iB log2(1 + β_iS_i), with technology-specific efficiency factors for LTE and Wi-Fi.Wi-Fi efficiency derives from CSMA/CA Markov-chain analysis, including empty, successful-transmission, and collision intervals.
  • II. BACKGROUND ON WI-FI/LTE CO-EXISTENCE: Wi-Fi suppresses transmission when detected channel energy exceeds its CCA threshold; otherwise, it transmits at a rate determined by SINR.This mechanism connects LTE energy at the Wi-Fi receiver to Wi-Fi throughput.

1) Characterization of Wi-Fi Throughput:

The Wi-Fi throughput characterization model uses Wi-Fi transmit power, channel gains, LTE interference, noise, and the CCA threshold to determine throughput with or without LTE.

  • 1) Characterization of Wi-Fi Throughput:: The Wi-Fi model takes Wi-Fi and LTE transmit powers, channel gains, noise power, and channel energy as inputs.The CCA threshold determines whether LTE energy prevents Wi-Fi transmission.
  • 1) Characterization of Wi-Fi Throughput:: The model outputs Wi-Fi throughput separately for operation without LTE and when LTE is present.The LTE-present case accounts for interference at the Wi-Fi receiver.

2) Characterization of LTE Throughput:

The LTE throughput characterization model uses LTE and Wi-Fi link parameters together with Wi-Fi channel energy to determine LTE throughput with or without Wi-Fi interference.

  • 2) Characterization of LTE Throughput:: LTE throughput is modeled using instantaneous rate updates based on the interference caused by Wi-Fi activity.The characterization follows the Wi-Fi active-time fraction produced by CSMA/CA.
  • 2) Characterization of LTE Throughput:: The LTE model takes LTE and Wi-Fi transmit powers, link gains, noise power, and Wi-Fi channel energy as inputs.The model also uses the Wi-Fi CCA threshold as a parameter describing channel activity.
  • 2) Characterization of LTE Throughput:: The model produces LTE throughput for operation without Wi-Fi and when Wi-Fi is present.The two cases distinguish standalone LTE from LTE affected by Wi-Fi transmission.

B. Experimental Validation

Experiments on ORBIT using USRP-based LTE and commercial Wi-Fi equipment evaluate interference as the distance between an interfering access point and the Wi-Fi link changes. The results broadly track the analytical model and show severe baseline degradation from LTE interference.

  • B. Experimental Validation: The validation uses an 802.11g Wi-Fi link with Atheros adapters and OpenAirInterface LTE on USRP platforms in the ORBIT testbed.LTE uses a 3GPP Release 8.6-compliant software implementation in SISO transmission mode.
  • B. Experimental Validation: The experiment varies the interfering AP distance from 1 to 20 m while measuring a Wi-Fi link with a fixed 0.25 m AP–client distance.Interference sources include LTE, another Wi-Fi link, and white noise on the same 2.4 GHz channel.
  • B. Experimental Validation: Analytical and experimental results closely match the trend of LTE’s effect on Wi-Fi as the interfering eNB distance varies.Some discrepancies are attributed to the fixed indoor environment and limited experimental data.
  • B. Experimental Validation: Even lightly loaded LTE control signals without scheduled LTE data transmission drastically reduce Wi-Fi performance.The analytical model incorporates the low-power LTE control-signal condition.
  • B. Experimental Validation: Up to 90% Wi-Fi throughput degradation occurs with uncoordinated LTE interference relative to interference-free Wi-Fi.LTE at 5 or 10 MHz causes 20–80% degradation relative to Wi-Fi interference, while other Wi-Fi reduces throughput approximately by half.

C. Motivational Example

The motivational example evaluates Wi-Fi and LTE throughput under co-channel interference while varying link and interferer distances. LTE interference severely degrades Wi-Fi, while Wi-Fi interference also reduces LTE throughput but with a smaller average impact.

  • C. Motivational Example: The setup evaluates a single associated Wi-Fi or LTE link against an interfering AP under co-channel operation.The experimental scenario includes Wi-Fi, LTE, other Wi-Fi, and white-noise interference cases in the 2.4 GHz band.
  • C. Motivational Example: The experiment varies associated-link distance from 0 to 100 m and interferer distance from −100 to 100 m in a horizontal deployment.The distance convention captures how Wi-Fi–LTE AP separation affects Wi-Fi transmission or shutdown through CCA sensing.
  • C. Motivational Example: Wi-Fi throughput falls to zero in 80% of cases and degrades by an average 91% under co-channel LTE interference.Within approximately 20 m, Wi-Fi’s CCA mechanism shuts off the AP; low-SINR cases also produce zero throughput.
  • C. Motivational Example: LTE throughput is zero across 45% of the evaluated area and degrades by an average 65% under Wi-Fi interference.Wi-Fi’s CCA busy region can shut off Wi-Fi, allowing LTE to operate as if Wi-Fi were absent.

IV. SYSTEM ARCHITECTURE

The proposed system architecture uses logically centralized SDN coordination to exchange network information and control spectrum-sharing decisions across heterogeneous Wi-Fi and LTE networks. It models multiple networks, CSMA effects, hidden-node interference, SINR, and throughput requirements for coordinated resource management.

  • IV. SYSTEM ARCHITECTURE: The SDN architecture coordinates heterogeneous networks through a Global Controller and Regional Controllers.The Global Controller processes network-state information and manages information flow among Regional Controllers and databases.
  • IV. SYSTEM ARCHITECTURE: The coordination framework forwards power and time-division channel-access parameters to networks according to UE throughput requirements.Both Wi-Fi and LTE share one channel and the same bandwidth, while clients remain associated with their original APs.
  • V. SYSTEM MODEL: The system model represents N Wi-Fi and M LTE networks and uses throughput functions based on link-specific SINR.The notation absorbs bandwidth into α_i, yielding R_i = α_i log2(1 + β_iS_i).
  • V. SYSTEM MODEL: Wi-Fi channel access is reduced by neighboring APs in CSMA range and by hidden nodes outside carrier-sensing range.The modeled access reductions are approximately 1/(1 + |M_a_i|) and 1/(1 + ζ|M_b_i|), respectively, with ζ in [0.2, 0.6].
  • V. SYSTEM MODEL: The framework updates Wi-Fi activity and interference factors to represent CSMA/CA’s reduced active time in SINR and throughput calculations.Wi-Fi interference is scaled by a_k because each Wi-Fi link is active for only an approximate fraction of time.
  • IV. SYSTEM ARCHITECTURE: Inter-network coordination is formulated to satisfy minimum throughput requirements and guarantee requested service availability at each UE.The optimization is implemented in two stages.

VI. COORDINATION VIA JOINT OPTIMIZATION

The joint power-control framework maximizes aggregate Wi-Fi-plus-LTE throughput while enforcing power, SINR, and Wi-Fi clear-channel-assessment constraints. When feasibility requires relaxing LTE minimum-throughput requirements, some LTE links may be shut off to protect neighboring Wi-Fi devices.

  • VI. COORDINATION VIA JOINT OPTIMIZATION: The optimization selects Wi-Fi and LTE transmit powers to maximize aggregate throughput across both networks.Power variables are bounded by P_max for both Wi-Fi and LTE APs within the SDN framework.
  • VI. COORDINATION VIA JOINT OPTIMIZATION: Aggregate throughput maximization is converted into maximizing a product of Wi-Fi and LTE SINR values under periodically updated Wi-Fi parameters.The formulation treats α_i and β_i as network constants and uses a high-SINR throughput approximation.
  • VI. COORDINATION VIA JOINT OPTIMIZATION: Each link must meet a minimum SINR requirement, while Wi-Fi additionally must satisfy a clear-channel-assessment threshold.The CCA constraint accounts for LTE interference, Wi-Fi interference within the interference zone, and noise at the Wi-Fi AP.
  • VI. COORDINATION VIA JOINT OPTIMIZATION: To preserve feasibility, the formulation can relax LTE minimum-throughput constraints to zero, effectively shutting off interfering LTE links.This relaxation can cause throughput deprivation at some LTE links.

B. Joint Time Division Channel Access Optimization

Because joint power control may deprive some LTE links of throughput, the paper proposes time-division channel access so Wi-Fi and LTE take turns using the shared channel. The objective is to maximize the minimum throughput across both networks.

  • B. Joint Time Division Channel Access Optimization: Time-division access is introduced because joint power control is insufficient when every UE must receive non-zero throughput.Power optimization can relax LTE requirements and deprive selected LTE links.
  • B. Joint Time Division Channel Access Optimization: The time-sharing variable η determines the fraction of channel access assigned to one RAT, with the other RAT receiving the remaining fraction.The formulation constrains η to [0, 1].
  • B. Joint Time Division Channel Access Optimization: The time-division objective maximizes the minimum throughput across Wi-Fi and LTE networks.The optimization is proposed in two steps.

1) Power control optimization across network of same RAT:

The formulation jointly optimizes Wi-Fi and LTE transmission powers under throughput, SINR, and power constraints, while time-division optimization maximizes minimum user throughput using prior throughput values.

  • Power control optimization: The power-control formulation optimizes access-point transmission powers for each RAT subject to minimum-rate and maximum-power constraints.The objective is equivalent to maximizing the product of network SINRs while satisfying link requirements.
  • Power control optimization: The formulation explicitly models Wi-Fi and LTE SINRs and imposes separate feasibility constraints for links in each network.The supplied constraints require minimum rates and bound transmission powers for Wi-Fi and LTE links.
  • Evaluation visuals: The section presents heat-map and feasibility-region visuals for Wi-Fi and LTE throughput under coordinated power and time-division access.The listed visuals include Wi-Fi and LTE throughput heat maps, feasibility regions, and time-division coordination results.
  • Joint time division channel access optimization: Joint time-division optimization fixes prior throughput values and chooses η to maximize the minimum throughput across all UEs.This optimization is performed after the preceding step and uses its throughput outputs as constants.

VII. EVALUATION OF JOINT COORDINATION

Single-link evaluations compare uncoordinated operation with joint power control and time-division coordination. Coordination improves throughput broadly, while time division removes infeasible topological regions and improves low-percentile performance and fairness.

  • Single-link results: Joint power control improves overall throughput for most topologies but leaves infeasible regions when close UE and interfering-AP placements violate Wi-Fi CCA and link-SINR requirements.Time-division coordination eliminates the infeasible region for the evaluated scenarios.
  • Single-link results: 15–20 Mbps 10-percentile throughput is achieved for both Wi-Fi and LTE with time-division coordination, versus approximately zero without or with power coordination.The comparison is reported for the single-link deployment summarized in Figure 12.
  • Single-link results: Wi-Fi mean throughput gains reach 200% with power coordination and 350% with time-division coordination relative to no coordination.LTE gains are approximately 25–30% for both coordination approaches.
  • Single-link results: Time-division coordination provides throughput fairness between Wi-Fi and LTE, although it offers LTE no additional advantage over power coordination.The reported fairness benefit concerns coexistence in the shared band.

B. Multiple Links Co-channel Deployment

Multi-link experiments average throughput across random deployments with equal numbers of Wi-Fi and LTE links. Uncoordinated LTE interference starves low-percentile Wi-Fi users, while coordination reallocates resources toward fairness and can favor orthogonal allocation as density grows.

  • B. Multiple Links Co-channel Deployment: 10-percentile Wi-Fi users become throughput-starved under no coordination because of LTE interference in multi-link deployments.The multi-link results are averaged over ten random topologies for N = 2, 5, and 10 links per network.
  • B. Multiple Links Co-channel Deployment: Coordination reallocates spectrum toward affected Wi-Fi links, introducing fairness while reducing LTE throughput relative to uncoordinated operation.Without coordination, LTE experiences no Wi-Fi interference because Wi-Fi defers transmission under LTE’s impact.
  • B. Multiple Links Co-channel Deployment: As the number of links grows, coordinated orthogonal resource allocation provides greater benefit than power-control optimization that fully shares the spectrum.For small link counts, joint time-division access degrades both Wi-Fi and LTE performance.
  • VIII. CONCLUSION: The conclusion reports that logically centralized coordination improves aggregate throughput and makes the networks’ aggregate throughputs comparable with joint power control and time division.Comparable aggregate throughput is presented as realizing fair spectrum access.
Loading 1507.06881v1…