Source-linked AI summary
Cooperative Dynamic Spectrum Access for Cell-Free MISO Networks Using Spectrum Consumption Models
Jiwon Sung, Seungmin Choi, Abhiram R. Gorle, Atul A. Salvekar, Igor Kadota, John M. Cioffi
TL;DR
SCM-based DSA has generally assumed fixed radiation patterns, limiting the design space for multiuser spectrum sharing. This paper jointly designs precoding and spectrum allocation for a cell-free UPA network, and its simulations show improved use of spatial dimension for sharing.
Problem
Prior SCM-based DSA assumes fixed radiation patterns, while phased arrays can adapt beams for multiuser spectrum sharing.
Source: §I · p. 1Method
The paper combines UPA-based precoding, leakage-aware iterative water-filling, SINR-targeted power control, and sequential SCM compatibility-test admission in cooperative DSA.
Source: §I · p. 1, §III · p. 3Results
The proposed mechanism better utilizes the spatial dimension for spectrum sharing; Voronoi-partition power control reduces spectrum use to roughly 2.6 slots but trades off spectral efficiency.
Source: §IV · pp. 5–6, §V · p. 6Takeaways & Limitations
Jointly designing the precoder and spectrum allocation is critical for spectrum sharing in the studied warehouse setting.
Source: §V · p. 6
Abstract
from arXiv · showhide
This paper incorporates uniform planar arrays into a cooperative dynamic spectrum access (DSA) mechanism based on the IEEE 1900.5.2 spectrum consumption model (SCM). Unlike prior work on SCM-based DSA that assume a fixed radiation pattern, the proposed DSA mechanism treats the radiation pattern as a design variable and jointly designs the precoder and spectrum allocation. The system model is a cell-free downlink, multiuser, multiple-input single-output network, an architecture that suits the dense indoor deployments of large-scale networks such as online retailers' fulfillment centers. The main objective is to reduce spectrum use while meeting a target signal-to-interference-plus-noise ratio at every receiver. To this end, this paper proposes a leakage-aware iterative water-filling precoder using uplink-downlink reciprocity, a power-control method that further encourages spectrum sharing, and a sequential admission algorithm that allocates spectrum to base stations based on the SCM compatibility test. Simulations in an indoor warehouse setting show that the proposed DSA mechanism better utilizes the spatial dimension for spectrum sharing.
I. INTRODUCTION
The paper extends SCM-based cooperative DSA to a cell-free multiuser system by making UPA radiation patterns design variables and jointly designing precoding and spectrum allocation. It situates this approach in warehouse-scale spectrum sharing and the IEEE compatibility-test framework.
I. INTRODUCTION
- Spatial beamforming can serve user groups on shared time-frequency resources, extending DSA beyond fixed-pattern directional links.The introduction motivates this capability for large-scale, time-sensitive deployments such as fulfillment warehouses.Source: §I · p. 1
- The proposed cooperative DSA jointly designs precoding and spectrum allocation, treating UPA radiation patterns as variables rather than fixed patterns.Its contributions include leakage-aware iterative water-filling, SINR-targeted power control, and sequential admission based on the compatibility test.Source: §I · p. 1
A. Background on Spectrum Consumption Models
- SCMs describe RF devices’ spectral, spatial, and temporal characteristics, and the compatibility test evaluates whether multiple devices can share spectrum without severe interference-driven degradation.The paper’s SCM uses transmitter and receiver constructs, including spectrum and underlay masks.Source: A. Background on Spectrum Consumption… · p. 2
II. SYSTEM MODEL
- The system model places B ceiling-mounted base stations with UPAs over a square coverage area and assigns uniformly located users to them.The BSs occupy a square grid; each user is served by one BS, and each BS’s UPA faces users below.Source: §II · p. 2
A. UPA Array Response Vector
The system represents each base station’s UPA using its geometry and an array response toward users. Users are divided into disjoint groups, each served by one base station at a time.
A. UPA Array Response Vector
- Each base station’s UPA has N=NxNy elements with half-wavelength spacing, and the array response toward a user uses the direction cosines.The text notes that this spacing is approximately consistent across carrier frequencies when occupied bandwidth is negligible relative to carrier frequency.Source: A. UPA Array Response Vector · p. 2
B. User Partitioning
- The partition assigns each user to one base station and divides users into disjoint groups served by their respective base stations.The baseline schemes are random balanced, greedy balanced, and Voronoi partitions; a base station with no assigned users is inactive.Source: B. User Partitioning · p. 2
C. Downlink Signal Model
Each base station transmits a precoded superposition for its assigned users. Received signals include the desired transmission and intra- and inter-group interference, with inter-base-station coupling governed by spectral overlap.
C. Downlink Signal Model
- Each base station transmits a precoded superposition of symbols for its assigned user group, using per-user precoding vectors and a base-station power value.The precoding matrix is formed from the user vectors and has an identity trace normalization.Source: C. Downlink Signal Model · p. 2
- Inter-base-station interference depends on spectral overlap: shared spectrum has overlap coefficient 1, while non-overlapping bands are treated as orthogonal in simulations.For frequency separation greater than the occupied passband bandwidth, the overlap coefficient is approximately zero.Source: C. Downlink Signal Model · pp. 2–3
- The received signal comprises the desired signal, intra-group interference, and inter-group interference, alongside additive white Gaussian noise.The channel from each base station to a user and the noise term are specified in the model.Source: C. Downlink Signal Model · p. 3
D. Achievable Rate
The achievable-rate section defines interference powers and expresses ergodic sum spectral efficiency as an expectation over user locations and channel realizations. The subsequent design jointly selects precoders, carrier frequencies, and transmit powers.
D. Achievable Rate
- The model distinguishes intra-group and inter-group interference powers for evaluating the achievable rate.These interference components arise from transmissions within a user’s group and from other base-station groups.Source: D. Achievable Rate · p. 3
- The ergodic sum spectral efficiency is averaged over user locations and channel realizations.The system design jointly addresses the precoders, carrier frequencies, and power assignments.Source: D. Achievable Rate · p. 3
III. PROPOSED COOPERATIVE DSA SCHEME
The cooperative DSA scheme manages interference through spatial precoding and spectral deconfliction. Base stations are admitted sequentially, and a carrier is shifted when spatial processing alone does not pass the compatibility test.
III. PROPOSED COOPERATIVE DSA SCHEME
- The scheme uses spatial precoding to suppress intra-group interference and spectral deconfliction to assign each base station a carrier and power that keep interference below a threshold.The spectral layer accounts for inter-group interference.Source: §III · p. 3
- Base stations and their user groups enter one at a time; each newly admitted station beams toward its group while protecting already-admitted users sharing its carrier.Those protected users form the victim set.Source: §III · p. 3
- If spatial processing does not pass the compatibility test, the base station shifts to the next spectrum slot and repeats the test.The spectral deconfliction layer follows the spatial layer.Source: §III · p. 3
A. Iterative Water-filling via Uplink-Downlink Reciprocity
The method recasts each base station’s downlink group as a virtual uplink using reciprocity, then iteratively updates user powers and interference-plus-noise covariance. Its leakage-aware regularization steers beams away from protected users.
A. Iterative Water-filling via Uplink-Downlink Reciprocity
- Uplink-downlink reciprocity converts each base station’s intra-group MISO downlink into a virtual-uplink SIMO multiple-access channel with the same channel vectors and a per-station power constraint.Inter-group interference mitigation is deferred to spectral deconfliction.Source: A. Iterative Water-filling via… · p. 3
- A leakage-aware regularization term, weighted by ζ, controls how strongly a base station steers beams away from users in its victim set.The resulting covariance includes the user’s own channel-and-power term.Source: A. Iterative Water-filling via… · p. 3
- The algorithm’s inversion matrix is Hermitian positive definite, and Algorithm 1 converged within nmax iterations in all simulations.The paper outlines the overall iterative procedure, including determination of the water level.Source: A. Iterative Water-filling via… · p. 4
- The IWF method alternates water-filling power updates with covariance updates because each depends on the other, and weighted-averages the resulting update.The water-filling solution satisfies the stated Karush-Kuhn-Tucker conditions with the covariance term held fixed.Source: A. Iterative Water-filling via… · p. 4
B. Precoding via Uplink-Downlink Reciprocity
At convergence, the virtual-uplink receive filter supplies the downlink beam direction, and the IWF precoder uses interference-plus-noise covariance to account for other users. MRT and RZF serve as comparison baselines.
B. Precoding via Uplink-Downlink Reciprocity
- At convergence, each user’s virtual-uplink linear minimum mean-square-error receive filter also serves as its downlink beam direction.The same vector is used for both roles.Source: B. Precoding via Uplink-Downlink… · p. 4
- The IWF precoder uses interference-plus-noise covariance to null more strongly toward users receiving large interference.The base-station precoder is normalized to unit total power.Source: B. Precoding via Uplink-Downlink… · p. 4
- MRT prioritizes users with the best channels without considering interference, whereas RZF prioritizes suppression of intra-group interference.Both baselines are normalized to unit power, and each base station scales power so its weakest-beamforming-gain user reaches the target SINR, up to Pmax.Source: B. Precoding via Uplink-Downlink… · p. 4
C. Downlink Power Allocation and Power Control
The power-control method limits each base station to the power its users need to reach the target SINR, subject to a maximum. The SCM underlay mask bounds aggregate interference, and reduced transmit power reduces leakage to neighboring groups.
C. Downlink Power Allocation and Power Control
- Each base station uses only the transmit power its users require to reach target SINR γtgt, capped at Pmax.The interference allowance IRx is the largest aggregate interference permitted by the SCM underlay mask; limiting power reduces leakage into neighboring groups.Source: C. Downlink Power Allocation and Power… · p. 4
D. SCM Compatibility Test
The IEEE 1900.5.2 compatibility test checks whether spectrum sharing keeps interference at victim users below their allowed thresholds. Sequential admission applies this test as base stations are considered for spectrum slots.
D. SCM Compatibility Test
- The admission procedure builds precoding weights and sets each base station’s power and spectrum allocation before testing compatibility.It defines a victim set and admits the base station when the test passes or no base station has yet been admitted.Source: D. SCM Compatibility Test · p. 5
- The compatibility test compares received interference power with the allowed threshold at every frequency for each victim user.A spectrum mask describes relative transmitted power by frequency offset, while an underlay mask describes the interference a user can tolerate.Source: D. SCM Compatibility Test · pp. 4–5
- If compatibility fails even on a fresh slot, the failure is attributed to intra-group interference, but the base station is still admitted on that slot.The test passes only when all admitted users are compatible during sequential admission.Source: D. SCM Compatibility Test · p. 5
IV. NUMERICAL RESULTS
In the warehouse simulations, precoding and power control affect spectrum reuse and spectral efficiency. IWF achieves the highest sum spectral efficiency per occupied slot across the tested user counts, while power control reduces slot use with the Voronoi partition.
Simulation setup
- The default parameters specify a 28 GHz carrier, 16 ceiling-mounted base stations, 64 users, and 64 array elements (8×8); results are averaged over 100 independent Monte Carlo trials.The target SINR and interference-to-noise ratio are both 6 dB.Source: §IV · p. 5
Spatial power distribution
- The spatial power map shows each base station concentrating emissions on its own group’s users, which facilitates spectrum sharing.It maps received power relative to the noise floor across the user-height plane after one IWF run with Voronoi partitioning.Source: §IV · p. 5, Figure 2 · p. 5
Power control
- Roughly 2.6 spectrum slots are used with power control and the Voronoi partition, versus nearly all available slots without power control.Without power control, the 16 base stations use maximum power and occupy nearly all slots with no reuse.Source: §IV · p. 5, Figure 3 · p. 6
- The Voronoi partition is the only scheme whose sum spectral efficiency per slot improves with power control; random and greedy partitions lose more throughput than slots saved.Power control prioritizes spectrum sharing over using maximum power.Source: §IV · pp. 5–6
Precoding comparison
- MRT uses more spectrum slots than RZF and IWF, while IWF also accounts for inter-group interference through leakage-aware regularization.RZF and IWF steer beams to reduce intra-group interference; the comparison varies the number of users.Source: §IV · p. 6, Figure 4 · p. 6
- IWF achieves the highest sum spectral efficiency per occupied slot at every tested user count.The comparison covers IWF, RZF, and MRT as K varies from 8 to 128.Source: §IV · p. 6, Figure 4 · p. 6
V. CONCLUSION
The paper concludes that jointly designing precoding and spectrum allocation is critical for better use of the spatial dimension in spectrum sharing.
V. CONCLUSION
- Jointly designing the precoder and spectrum allocation is critical for better use of the spatial dimension in spectrum sharing.The cooperative SCM-based DSA mechanism uses uniform planar arrays at the transmitter.Source: §V · p. 6