Spectrum Energy Balance Coordinated for Unmanned Systems by Channel Interference in Low-Altitude Intelligent Network

0
0
0

Spectrum Energy Balance Coordinated for Unmanned Systems by Channel Interference in Low-Altitude Intelligent Network

Introduction

The rapid development of unmanned systems has led to increasing demands for high-density, large-traffic, and near-real-time low-altitude applications. To meet these requirements, low-altitude intelligent networks (LAIN) have emerged, forming a diverse, heterogeneous, and dynamic operational ecosystem. Among the critical challenges in such networks, balancing spectrum efficiency, energy efficiency, and network throughput has become a focal point, particularly in resource-constrained unmanned systems.

Interference management through power control is an effective method to optimize both energy and spectrum resources. By mitigating interference while reducing energy consumption, this approach enhances regional spectrum utilization and network throughput. Unmanned systems, often deployed in infrastructure-less environments, provide flexible wireless connectivity through broadcast or multicast line-of-sight (LoS) communication links. Their advantages include rapid deployment, reconfigurability, and improved communication performance. Future swarm-centric networks are expected to adopt hierarchical, distributed clustering structures as the primary networking paradigm.

This paper introduces a network throughput optimization method that integrates interference coordination and joint power control to address channel interference and power allocation challenges in unmanned systems. The proposed method maximizes overall network throughput while balancing energy and spectrum efficiency under Doppler shift and external interference conditions.

Network Architecture and Node Deployment

Group-Centric Network (GCN) Framework

Given the heterogeneous nature of unmanned platforms—spanning aerial, ground, and maritime domains—a unified group-centric network framework is essential for seamless integration and adaptability. The GCN architecture employs a two-tier hierarchical clustering structure:

  1. Intra-Cluster Communication Layer: This layer facilitates communication between cluster head nodes (CHNs) and cluster member nodes (CMNs) within sub-groups.
  2. Inter-Cluster Communication Layer: CHNs from different sub-groups communicate through a centralized processing node (CPN), which manages network operations.

The CPN, equipped with multiple-input multiple-output (MIMO) antennas, provides wireless communication services, while individual unmanned systems operate in full-duplex mode with single antennas.

Mobility Control

Unmanned systems adjust their positions based on CPN or CHN movement to maintain LoS communication. A random waypoint (RWP) mobility model is used to simulate dynamic exploration in three-dimensional space. Two control methods are employed:

  1. Centralized Management: The CPN controls CHNs, which in turn manage energy and frequency resources within their sub-groups.
  2. Distributed Optimization: In scenarios where CPN or CHN failures occur, local self-optimization ensures network resilience.

Node Information Transmission

High-speed, broadband data transmission is crucial for unmanned systems, especially in dense deployments with limited channel resources. Different full-duplex modes introduce interference challenges:

• Out-of-Band Full-Duplex (OBFD): Uplink and downlink transmissions occur on orthogonal channels.

• In-Band Full-Duplex (IBFD): Uplink and downlink share the same frequency band.

The CPN uses zero-forcing precoding to mitigate self-interference from MIMO transmissions. Additionally, the Clarke flat-fading channel model approximates Doppler shift effects, while external interference is also considered.

Interference Model in Low-Altitude Intelligent Networks

Given the limited availability of uplink and downlink channels, interference coordination is critical. Based on full-duplex operational modes, three interference types are analyzed:

  1. Cross-Layer Interference

When CHNs operate in OBFD mode, nearby CMNs experience cross-layer interference from CHNs transmitting on uplink channels. The received signal at a CMN includes contributions from the CPN, interfering CHNs, external interference, and noise. The signal-to-interference-plus-noise ratio (SINR) is derived to quantify performance degradation.

  1. Intra-Layer Interference

a) OBFD Mode
CHNs in OBFD mode suffer intra-layer interference from other CHNs transmitting on downlink channels. The SINR is influenced by CPN transmissions, neighboring CHN interference, and noise.

b) IBFD Mode
CHNs in IBFD mode experience additional self-interference from their own transmissions, further degrading SINR.

  1. Mixed-Layer Interference

a) OBFD Mode
CMNs in OBFD mode face combined cross-layer interference from the CPN and intra-layer interference from other CHNs.

b) IBFD Mode
CMNs in IBFD mode encounter interference from both the CPN and other CHNs sharing the same frequency band.

Joint Power Control and Interference Coordination

To maximize network throughput, the interference coordination problem is formulated as a combinatorial optimization problem. The objective function accounts for:

• Throughput Contributions: From CPN-to-CMN, CPN-to-CHN (OBFD), and CPN-to-CHN (IBFD) communications.

• Binary Variables: Representing CHN operational modes (0 for OBFD, 1 for IBFD).

• Continuous Variables: Adjusting CHN transmission power weights.

The optimization problem is non-convex, requiring specialized solution strategies.

Optimization Strategies

  1. Centralized Iterative Methods

a) Iterative Backtracking Optimization
The problem is decomposed into two sub-problems: power allocation (continuous) and mode selection (discrete). These are solved iteratively until convergence.

b) Iterative Search with Steepest Descent (IS-SD)
Gradient-based optimization is applied to both power and mode variables, with step sizes determined via line search.

  1. Distributed Intelligent Strategies

a) Independent Subspace Particle Swarm Optimization (IS-PSO)
Particle swarms explore power and mode subspaces independently, with dynamic inertia weights and adaptive memory to avoid local optima.

b) Double Genetic Algorithm (DGA)
Two populations evolve power and mode solutions through crossover and mutation operations.

c) Greedy Algorithm with Penalty Function (GRA-PF)
A penalty function enforces boundary conditions on power weights, guiding the greedy search toward feasible solutions.

Performance Evaluation

Simulation Setup

A network of 100 unmanned systems is simulated, with 10 sub-groups of equal size. Key parameters include:

• CPN coverage: 1000 feet

• CMN coverage: 100 feet

• Carrier frequency: 1.215 GHz

• Bandwidth: 20 MHz

• Maximum speed: 15 feet/s

Results

  1. Network Throughput vs. MIMO Antennas
    Throughput increases with antenna count but exhibits periodic fluctuations due to self-interference. IS-SD performs poorly in high-antenna configurations (22–32 antennas).

  2. IBFD Mode Adoption
    Higher MIMO antenna counts at the CPN increase IBFD adoption among CHNs, improving spectrum utilization.

  3. Power Allocation
    Energy-efficient solutions allocate lower power weights to CHNs, extending operational endurance.

  4. Computational Complexity
    IS-PSO and IS-SD offer faster convergence than DGA, making them suitable for large-scale deployments.

Conclusion

This paper presents a comprehensive framework for interference coordination and power optimization in low-altitude intelligent networks. By leveraging MIMO and full-duplex technologies, the proposed method maximizes throughput while balancing energy and spectrum efficiency. Centralized and distributed optimization strategies provide flexible solutions for varying network conditions. Future work will explore real-world implementations and scalability enhancements.

doi.org/10.19734/j.issn.1001-3695.2024.04.0142

Was this helpful?

0 / 0