PaperlayerView exact revision on arXiv

Source-linked AI summary

Cooperative Interference Mitigation and Handover Management for Heterogeneous Cloud Small Cell Networks

Haijun Zhang, Chunxiao Jiang, Julian Cheng, Victor C. M. Leung

arXiv:1504.08076v2cs.ITcs.NI

TL;DR

HCSNet must address co-channel interference and handover management, especially for cell-edge users. The paper combines C-RAN with small cells, applies affinity-propagation CoMP clustering, and presents handover management; numerical results show increased capacity while maintaining users’ quality of service.

  • Problem

    Cooperative interference mitigation and handover management in HCSNet remain insufficiently investigated, especially for cell-edge users.

  • Method

    The paper combines C-RAN with small cells, uses affinity propagation for CoMP clustering, and presents handover management with signaling analysis.

  • Results

    Numerical results show that the proposed architecture, CoMP clustering, and handover management increase HCSNet capacity while maintaining users’ quality of service.

  • Takeaways & Limitations

    The proposed combination supports cooperative interference mitigation and handover management in HCSNet while maintaining users’ quality of service.

Abstract

from arXiv · show

Heterogeneous small cell network has attracted much attention to satisfy users' explosive data traffic requirements. Heterogeneous cloud small cell network (HCSNet), which combines cloud computing and heterogeneous small cell network, will likely play an important role in 5G mobile communication networks. However, with massive deployment of small cells, co-channel interference and handover management are two important problems in HCSNet, especially for cell edge users. In this article, we examine the problems of cooperative interference mitigation and handover management in HCSNet. A network architecture is described to combine cloud radio access network with small cells. An effective coordinated multi-point (CoMP) clustering scheme using affinity propagation is adopted to mitigate cell edge users' interference. A low complexity handover management scheme is presented, and its signaling procedure is analyzed in HCSNet. Numerical results show that the proposed network architecture, CoMP clustering scheme and handover management scheme can significantly increase the capacity of HCSNet while maintaining users' quality of service.

I. INTRODUCTION

HCSNet combines cloud radio access networks with heterogeneous small cells to address traffic demand, while targeting interference and mobility challenges. The paper proposes CoMP-based interference mitigation and handover management for these issues.

  • Motivation: HCSNet combines C-RAN and heterogeneous small cells to reduce energy costs and facilitate centralized radio resource management.C-RAN centralizes baseband processing in BBU pools while distributing RF processing to RRHs.
  • Challenges: Dense small-cell deployment creates co-channel interference, particularly affecting cell-edge users.Cloud-enabled centralized processing facilitates interference mitigation techniques including eICIC and CoMP.
  • Research gap: Efficient, low-complexity CoMP clustering remains insufficiently investigated in HCSNet.The paper identifies clustering as an important factor affecting CoMP performance.
  • Research gap: Handover management in HCSNet has received little attention, and prior work lacked quantitative analysis.The paper focuses specifically on handover management rather than mobility management broadly.
  • Contributions: The paper presents an HCSNet architecture, an affinity-propagation CoMP clustering scheme, and a handover management scheme with analyzed signaling procedures.The proposed methods target cooperative interference mitigation and handover management.

II. HCSNET ARCHITECTURE

The HCSNet architecture co-locates macrocell and small-cell baseband processing in BBU pools while distributing radio functions across MRRHs and SRRHs. This centralization supports cooperative processing and handles frequent small-cell handovers.

  • Architecture: HCSNet replaces macro and small-cell base stations with macro RRHs and small RRHs connected to centralized BBU pools.BBU pools provide baseband processing, while RRHs provide wireless coverage from different sites.
  • Architecture: BBU pools connect through X2 interfaces and communicate with RRHs over fronthaul links.Millimeter-wave radio is used for fronthaul between BBU pools and RRHs.
  • Deployment: RRHs can be deployed across buildings, offices, and hotspots such as stadiums to enhance coverage and capacity.The architecture supports both distributed indoor deployment and hotspot scenarios.
  • Interference management: Centralized processing enables efficient cooperative interference management despite potentially high interference in hotspot deployments.The BBU-pool architecture reduces processing and transmission delay.
  • Mobility: High-mobility users may experience frequent handovers because small cells have limited coverage areas.Handover management is processed in the BBU pools.

III. COOPERATIVE INTERFERENCE MITIGATION USING COMP IN HCSNET

Small-cell deployment produces co-channel interference in HCSNet, making interference management essential. Centralized BBU-pool processing facilitates mitigation methods such as eICIC and CoMP.

  • Interference challenge: Dense small-cell deployment results in co-channel interference, so interference management is essential in HCSNet.The problem is especially relevant to the network’s heterogeneous deployment.
  • Centralized processing: Centralized signal processing in BBU pools reduces processing and transmission delay.This centralized processing supports interference mitigation in HCSNet.
  • Mitigation mechanisms: Enhanced inter-cell interference coordination and CoMP are facilitated by the HCSNet architecture.These techniques are presented as interference-mitigation mechanisms enabled by centralized processing.

A. Cloud CoMP Architecture for Interference Mitigation

Cloud CoMP uses RRH clusters to jointly serve users and mitigate cell-edge interference, but clustering must balance performance against signaling and information-exchange complexity. The paper adopts affinity propagation using limited local information and CSI.

  • Cloud CoMP architecture: CoMP mitigates cell-edge inter-cell interference by coordinating RRHs that share measurement information such as power levels and CSI.The paper focuses on joint transmission, where clustered RRHs jointly serve users.
  • Cloud CoMP architecture: A coordinated RRH cluster jointly receives and processes data for a CoMP user through cloud-connected fronthaul links.Control signaling and user data are exchanged over these links.
  • Clustering trade-offs: Static clustering is simple but provides limited throughput gain, whereas full-dynamic clustering can incur large signaling and information-collection overhead.Full-dynamic schemes can achieve good performance at the expense of exhaustive information interchange and collection.
  • Clustering trade-offs: Semi-dynamic clustering balances performance and complexity by selecting measurement clusters without dynamic channel information and coordinated clusters with it.The distinction separates measurement-cluster selection from coordinated-cluster selection.
  • Affinity-propagation clustering: Affinity propagation is used for CoMP RRH selection because complete CSI is difficult to obtain in densely deployed small cells.The proposed similarity matrix uses local information and limited CSI between neighboring RRHs.

B. AP Clustering Based Cloud CoMP Scheme in HCSNet

The paper presents a semi-dynamic CoMP clustering framework for HCSNet that balances performance and complexity through offline measurement clustering and online coordinated-cluster selection. Affinity propagation uses similarity information, including pair CoMP SINR gain, to form coordinated RRH clusters with limited cooperation and CSI.

  • Framework: The proposed semi-dynamic clustering framework uses offline and online phases to maximize spectrum efficiency and throughput with low fronthaul traffic.The offline phase forms measurement RRH clusters using geographical location and reference signal received power; the online phase selects coordinated RRHs.
  • Design tradeoff: The semi-dynamic scheme is presented as effective compared with static and full-dynamic clustering while requiring limited CSI between local and neighboring cells.Only a limited number of RRHs cooperate to keep communication overhead affordable.
  • Framework: Offline clustering identifies measurement RRH clusters, while online affinity propagation selects coordinated RRH clusters from them.The online phase uses periodically fed-back sounding-reference-signal SINR measurements.
  • Affinity propagation: APBC takes similarity as input, updates responsibility and availability, and outputs exemplars with their associated RRH nodes as coordinated clusters.An exemplar represents the master RRH of a cluster, and preference values affect exemplar selection.
  • Affinity propagation: Pair CoMP SINR gain defines the off-diagonal similarity values as the ratio of estimated CoMP to non-CoMP SINRs for an RRH pair.The similarity matrix is APBC’s unique input and directly affects performance.

C. Evaluation of AP Clustering Based Cloud CoMP Scheme in HCSNet

The evaluation compares APBC with other CoMP clustering schemes on edge-user spectrum efficiency and execution time. APBC provides higher edge-user spectrum efficiency and, for more than 19 small cells, the lowest run time among the considered schemes.

  • Spectrum efficiency: CoMP schemes achieve higher spectrum efficiency than non-CoMP, while APBC provides better edge-user spectrum efficiency than static CoMP, sim-CoMP, and Wesemann’s scheme.The paper attributes APBC’s advantage to its information-interchange mechanism and pair CoMP SINR gain input.
  • Run time: When the number of small cells is greater than 19, APBC CoMP has the lowest run time among the four considered schemes.Run time increases for all four schemes as the number of RRHs increases.
  • Complexity: APBC’s semi-dynamic clusters allow implementation using local information and limited CSI between neighboring RRHs.The paper links this design to higher throughput with lower computational complexity.
  • Complexity: Each affinity propagation iteration requires O(n3) for n RRHs, based on responsibility and availability updates.The reported complexity depends on the number of algorithm iterations.
  • Complexity: The affinity propagation algorithm’s convergence has been proven, which the paper states guarantees the practicality of the presented algorithm.

IV. HANDOVER MANAGEMENT IN HCSNET

Handover management is a key QoS concern in HCSNet because dense small-cell deployment produces frequent handovers and exposes users to signaling burden, radio link failure, and unnecessary handovers. C-RAN architecture can reduce handover interruption time and delay, while BBU-pool processing can complete handovers within the pool.

  • QoS: Handover management is important for satisfying users’ quality-of-service requirements in mobile communications.The paper also states that handover can be accomplished within the BBU pool.
  • Challenges: Frequent handovers in densely deployed HCSNet burden the fronthaul and core networks.
  • Challenges: Small coverage areas and severe co-channel interference can produce mobility-related radio link failure and unnecessary handovers such as ping-pong handover.
  • Architecture: C-RAN-enabled HCSNet can reduce handover interruption time and delay.

A. Handover Procedures in HCSNet

In HCSNet, mobility functions are moved from radio heads to BBU pools, making handover procedures differ from traditional LTE procedures. The paper introduces an inter-BBU-pool MRRH-SRRH handover call flow and focuses its analysis on that scenario.

  • Handover flow: MRRH-SRRH handover is more challenging than SRRH-SRRH handover because macrocells and small cells have different coverage sizes.
  • HCSNet procedures: Handover decision and admission control are moved from SRRHs and MRRHs to the BBU pool, which supports their mobility-management functions.The SRRHs and MRRHs do not support mobility-management functions themselves.
  • Handover flow: The paper introduces an inter-BBU-pool MRRH-SRRH handover call flow for the HCSNet architecture.Handover between SRRHs follows the MRRH-SRRH handover procedure.
  • Handover flow: The intra-BBU-pool signaling flow is simpler than the inter-BBU-pool flow.The article focuses on the inter-BBU-pool scenario because of space limitations.

B. Inappropriate Handover Detection Method in HCSNet

The method detects handover-related radio link failures (RLFs) in dense HCSNet deployments and categorizes them by the UE’s subsequent connection state. It addresses several forms of inappropriate handover, including ping-pong, continue, late, early, and wrong handovers.

  • Dense small-cell deployment increases the number of handover-related RLFs.
  • Ping-pong and continue handovers reconnect the UE shortly after success to the serving cell or another RRH, respectively.
  • Late and early handovers involve an RLF before or during handover, or shortly after success, followed by reconnection to the target or serving RRH.
  • The cloud detects these scenarios by timing each UE after handover completion and stopping the timer when another RRH reports an RLF.
  • The RRH categorizes each RLF as a call drop, too-late handover, too-early handover, or wrong handover according to the UE’s post-RLF status.

C. A Low Complexity Handover Optimization Scheme in HCSNet

The proposed low-complexity handover scheme uses user speed and service type to avoid unnecessary macrocell-to-small-cell handovers. Its optimized signaling overhead decreases with more high-mobility users and remains lower than the traditional scheme.

  • Handover policy: The policy reduces unnecessary handovers because traditional treatment of high- and low-speed users can cause two unnecessary handovers during macrocell-to-small-cell movement.
  • Handover policy: The scheme prevents high-speed users from handing over to small cells, while allowing low-speed users to do so.
  • Handover policy: Medium-speed users with real-time service hand over to small cells, whereas those with non-real-time service do not.
  • Implementation: UE speed can be estimated from Doppler spread frequency or autocorrelation.
  • Evaluation: The optimized scheme is evaluated against the traditional scheme using signaling overhead across different scenarios.
  • Signaling overhead: For both algorithms, signaling cost increases with average session holding time because longer sessions produce more cell crossings and expected handovers.
  • Signaling overhead: As the high-mobility proportion α increases, traditional signaling overhead increases, whereas optimized-scheme overhead decreases toward zero as α approaches 1.

V. CONCLUSION

The paper combines cloud radio access networks with small cells and applies affinity-propagation CoMP clustering plus a handover-management scheme for HCSNet. Numerical results report increased capacity while maintaining users’ quality of service, with several joint extensions left for future work.

  • The proposed HCSNet architecture combines cloud radio access networks with small cells.
  • Affinity-propagation CoMP clustering is used to mitigate interference for cell-edge users.
  • The paper presents a handover-management scheme and analyzes its handover signaling procedures in HCSNet.
  • The proposed architecture, CoMP clustering, and handover-management scheme can significantly increase HCSNet capacity while maintaining users’ quality of service.
  • Future work will consider joint interference mitigation and handover management, joint time delay and clustering, and self-organized HCSNet.
  • The paper also identifies self-optimizing power control, automatic neighbor relation, and physical cell ID self-configuration as ways self-organized HCSNet can enhance these functions.
Loading 1504.08076v2…