UAV-Assisted 3-D Localization for IoT Networks Using a Simple and Efficient TDOA-AOA Estimator

Basic Information

• Paper ID: 2510.08765
• Title: UAV-Assisted 3-D Localization for IoT Networks Using a Simple and Efficient TDOA-AOA Estimator
• Authors: Mojtaba Amiri (University of Tehran), Rouhollah Amiri (Sharif University of Technology)
• Classification: eess.SP (Signal Processing)
• Publication Date: October 9, 2025 (arXiv Preprint)
• Paper Link: https://arxiv.org/abs/2510.08765

Abstract

This paper proposes an algebraic solution for three-dimensional source localization using minimal measurement data—specifically, one Time Difference of Arrival (TDOA) and one Angle of Arrival (AOA) pair. The method employs a closed-form weighted least squares (WLS) estimator that achieves localization using only a single ground station and one cooperative unmanned aerial vehicle (UAV) relay. Theoretical derivations and simulation results demonstrate the effectiveness of the proposed method, achieving near-optimal performance approaching the Cramér-Rao Lower Bound (CRLB) under moderate Gaussian noise conditions.

Research Background and Motivation

Core Problem

This research addresses three-dimensional source localization in Internet of Things (IoT) networks, particularly in scenarios where GNSS signals are unavailable or unreliable, aiming to achieve high-precision localization using minimal measurement data.

Significance Analysis

1. IoT Application Requirements: Localization plays a fundamental role in the IoT ecosystem, supporting applications such as intelligent transportation, asset tracking, environmental monitoring, and surveillance
2. Infrastructure Limitations: Traditional methods require multiple sensor devices, making implementation on lightweight aerial or space IoT platforms (such as UAVs and CubeSats) challenging
3. Synchronization Complexity: Existing methods typically require complex clock synchronization and hardware constraints

Limitations of Existing Methods

1. High Hardware Requirements: Existing joint TDOA-AOA methods require at least two sensors with angle measurement capabilities, necessitating antenna arrays
2. High Complexity: Multi-stage estimators increase computational complexity and synchronization requirements
3. Excessive Measurement Data: Traditional methods require more TDOA and AOA measurement pairs

Research Motivation

To propose a simple and efficient scheme requiring only a single ground sensor and one UAV relay, significantly reducing hardware and synchronization complexity while maintaining high localization accuracy.

Core Contributions

1. Minimal Measurement Scheme: Proposes a three-dimensional localization method using only one TDOA and one AOA measurement pair, representing the minimum measurement requirement in this field
2. Closed-Form Solution: Develops a closed-form estimator based on weighted least squares, avoiding the complexity of iterative optimization
3. Hardware Simplification: Achieves a localization architecture requiring only a single ground station and one cooperative UAV, substantially reducing hardware requirements
4. Theoretical Performance Guarantee: Theoretical analysis demonstrates that the method can achieve CRLB performance under low-noise conditions
5. Practical Validation: Simulation verification confirms the method's effectiveness under various noise conditions and target locations

Methodology Details

Task Definition

Input:

• Ground sensor position s₁ = x₁, y₁, z₁ᵀ
• UAV position s₂ = x₂, y₂, z₂ᵀ
• TDOA measurement r
• AOA measurement (azimuth φ, elevation angle θ)

Output:

• Estimated unknown source position û = x̂, ŷ, ẑᵀ

Constraints:

• Measurement noise follows zero-mean Gaussian distribution
• Noise covariance matrix is known

Model Architecture

1. Problem Formulation

The true TDOA measurement is defined as:

r⁰ = ‖u⁰ - s₂‖ - ‖u⁰ - s₁‖

The true AOA measurement is defined as:

φ⁰ = atan((y⁰ - y₁)/(x⁰ - x₁))
θ⁰ = atan((z⁰ - z₁)/((x⁰ - x₁)cos(φ⁰) + (y⁰ - y₁)sin(φ⁰)))

Measurement vector: m = r, φ, θᵀ = m⁰ + Δm

2. Pseudo-Linearization Processing

Through mathematical transformation, the nonlinear TDOA equation is converted to pseudo-linear form:

From the differential distance equation:

r⁰ + ‖u⁰ - s₁‖ = ‖u⁰ - s₂‖

After squaring:

r⁰² + ‖s₁‖² - ‖s₂‖² + 2r⁰‖u⁰ - s₁‖ + 2u⁰ᵀ(s₂ - s₁) = 0

3. AOA Equation Linearization

Utilizing geometric relationships, AOA measurements are converted to linear constraints:

αᵀs₁ - αᵀu⁰ = 0
βᵀs₁ - βᵀu⁰ = 0

Where:

• α = sin(φ⁰), -cos(φ⁰), 0ᵀ
• β = sin(θ⁰)cos(φ⁰), sin(θ⁰)sin(φ⁰), -cos(θ⁰)ᵀ

4. Unified Linear System

Combining TDOA and AOA equations into a unified linear system:

h = Gu⁰

Where:

• h = hᵣ, hφ, hθᵀ
• G = gᵣ, α, βᵀ

Technical Innovations

1. Single-Stage Estimator Design

Unlike traditional multi-stage WLS methods, this work leverages AOA data to eliminate nuisance parameters in the pseudo-linear equation, achieving single-stage estimation and avoiding error propagation.

2. Adaptive Weighting Strategy

A weighting matrix dependent on target position is designed:

W = (BQ_mBᵀ)⁻¹

Optimal weighting is achieved through iterative updates.

3. Closed-Form Solution

The final WLS estimator is:

û = (G̃ᵀWG̃)⁻¹G̃ᵀWh̃

This solution avoids the complexity and convergence issues of nonlinear optimization.

Experimental Setup

Simulation Parameters

• Ground Sensor Position: s₁ = 0, 0, 0ᵀ m
• UAV Position: s₂ = 500, 100, 2000ᵀ m
• Monte Carlo Trials: 10,000
• Noise Model: Zero-mean Gaussian distribution

Performance Metrics

• Root Mean Square Error (RMSE): RMSE(u) = √(Σ|u⁽ˡ⁾ - u⁰|²/L)
• Cramér-Rao Lower Bound (CRLB): Serves as theoretical performance benchmark

Experimental Scenarios

1. Scenario 1: Fixed target position with varying noise levels
2. Scenario 2: Target moving along x-axis with fixed noise level
3. Scenario 3: Performance comparison with Maximum Likelihood (ML) estimator

Experimental Results

Main Results

Scenario 1: Noise Sensitivity Analysis

• Target Position: u⁰ = 1000, 200, 100ᵀ m
• Noise Parameters: σᵣ = 40ρ, σφ = σθ = 0.1ρ
• Result: Across a wide range of noise conditions, the RMSE of the proposed method aligns closely with the CRLB

Scenario 2: Position Robustness Testing

• Noise Setting: (σᵣ, σφ, σθ) = (10m, 1°, 1°)
• Result: For all target positions within the monitoring region, the method achieves CRLB performance

Scenario 3: Comparison with ML Estimator

• Noise Setting: (σᵣ, σφ, σθ) = (10m, 1°, 1°)
• Result: The localization performance of the proposed method is comparable to the ML estimator, with significantly reduced computational complexity

Performance Analysis

1. Theoretical Optimality: Under low-noise conditions, the covariance matrix approaches the CRLB
2. Unbiased Estimation: Theoretical analysis proves the estimator is unbiased
3. Computational Efficiency: The closed-form solution avoids the computational burden of iterative optimization

Experimental Findings

1. The method exhibits excellent performance under moderate noise conditions
2. Hardware requirements are substantially reduced compared to traditional methods
3. Stable performance is maintained across different target positions

Related Work

Traditional Localization Methods

1. TOA Method: Requires precise clock synchronization
2. TDOA Method: Eliminates synchronization with the transmitter but typically requires multiple sensors
3. AOA Method: Determines direction through triangulation but suffers from geometric dilution of precision
4. RSS Method: Based on signal strength but is significantly affected by environmental factors

Hybrid TDOA-AOA Methods

Existing research, such as work by Yin et al. (2015) and Jia et al. (2018), requires at least two sensors with angle measurement capabilities, increasing hardware complexity.

Advantages of This Work

Compared to existing methods, this paper achieves minimal measurement requirements and hardware configuration while maintaining high-precision performance.

Conclusions and Discussion

Main Conclusions

1. Successfully achieves three-dimensional localization using only one TDOA and one AOA measurement pair
2. The developed closed-form WLS estimator can theoretically achieve CRLB performance
3. Significantly simplifies hardware requirements and synchronization complexity
4. Simulation verification confirms the method's effectiveness under various conditions

Limitations

1. Noise Assumptions: The method assumes Gaussian noise distribution; non-Gaussian noise may exist in practical environments
2. Low-Noise Conditions: Theoretical optimality is guaranteed only under low-noise conditions
3. Static Assumption: The current method assumes fixed sensor and UAV positions
4. Single-Source Limitation: Applicable only to single-source localization

Future Directions

1. Extension to multi-source localization scenarios
2. Consideration of non-Gaussian noise environments
3. Investigation of dynamic UAV trajectory optimization
4. Verification on actual hardware platforms

In-Depth Evaluation

Strengths

1. Strong Innovation: First to achieve three-dimensional localization with minimal measurement configuration
2. Rigorous Theory: Provides comprehensive theoretical analysis and performance guarantees
3. High Practical Value: Significantly reduces deployment costs of IoT localization systems
4. Simple Methodology: Closed-form solution avoids complex optimization processes
5. Sufficient Verification: Multi-scenario simulations validate method effectiveness

Weaknesses

1. Limited Application Scope: Verified only under moderate noise conditions; performance in high-noise environments unknown
2. Experimental Limitations: Lacks verification on actual hardware platforms
3. Insufficient Comparisons: Due to problem novelty, lacks direct comparison with similar methods
4. Environmental Assumptions: Assumes ideal propagation environment without considering multipath effects and other practical factors

Impact

1. Academic Value: Provides new research directions for TDOA-AOA localization theory
2. Practical Significance: Offers feasible localization solutions for resource-constrained IoT systems
3. Technology Promotion: Method is extensible to other wireless localization applications
4. Industrial Application: Promising for deployment in UAV-assisted IoT networks

Applicable Scenarios

1. UAV-Assisted IoT Networks: Particularly suitable for rapidly deployable temporary networks
2. Resource-Constrained Environments: Applicable to scenarios with limited hardware and energy resources
3. Emergency Response: Enables rapid establishment of localization networks for search and rescue operations
4. Environmental Monitoring: Suitable for large-scale environmental sensor network localization requirements

References

The paper cites 20 relevant references covering important works in localization theory, signal processing, UAV applications, and other related fields, providing a solid theoretical foundation for the research.

Overall Assessment: This is a paper with significant contributions to the IoT localization field. The proposed method is theoretically novel and practically valuable, providing an effective solution for resource-constrained localization applications. Despite certain limitations, its innovation and practicality confer substantial academic and application value.