Opportunity
The rapid growth of mobile communication demands, particularly with the advent of 5G technology, has led to an explosion in the number of connected devices and data transmission volumes. A significant challenge arises in indoor environments, where approximately 80% of data traffic occurs. 5G high-frequency signals suffer from severe attenuation over distance and when penetrating buildings, making it difficult for outdoor base station antennas to provide high-quality, comprehensive coverage indoors. In many indoor scenarios, such as offices, shopping malls, underground parking lots, and elevators, 5G signal coverage can be less than 60%, and even below 50% in more confined spaces. This creates a critical gap in achieving ubiquitous 5G connectivity. Indoor Distributed Antenna Systems (IB-DAS) are widely deployed to enhance mobile services inside buildings. However, the existing methods for planning antenna layouts within these systems are inefficient and inadequate. Current 5G network planning software primarily focuses on outdoor base station distribution and often employs single-objective optimization, failing to provide optimal trade-off solutions between competing objectives like construction cost (materials,工期) and network quality (power loss, signal coverage). Furthermore, existing indoor antenna layout optimization methods suffer from low algorithmic efficiency, often outputting only one result per computation, and rely on inaccurate signal propagation models unsuitable for simulating 5G high-frequency signals, leading to significant deviations between optimized and actual performance. Methods using accurate propagation models, like ray-tracing, are computationally expensive, making the optimization process costly and impractical for large-scale or complex indoor deployments.
Technology
This patent presents a novel method for antenna layout planning in indoor distributed systems, specifically designed for 5G networks. The core innovation lies in formulating the antenna placement problem as an expensive multi-objective optimization challenge and applying a decomposition-based Bayesian optimization framework to solve it efficiently. The method integrates several advanced techniques. It utilizes a ray-tracing propagation model to accurately simulate 5G signal strength within indoor environments, addressing the inaccuracy of simpler models. To manage the high computational cost of ray-tracing, the method employs a distribution prediction surrogate model, specifically a Gaussian process model, to approximate the expensive-to-evaluate objective functions, such as signal coverage rate, signal interference rate, and layout cost. The algorithm encodes potential antenna layout schemes into a real-number represented population. It uses a genetic algorithm with differential evolution crossover and mutation operators to generate offspring populations. A decomposition-based environmental selection strategy is then applied to select the best candidate population from the combined parent and offspring groups. Crucially, in each iteration, the method selects a small number (K_E) of representative individuals from the candidate population based on an "Expected Improvement" criterion calculated using the surrogate model's predictions and uncertainty estimates. These selected individuals undergo "real evaluation" using the accurate but costly ray-tracing model, and their results are used to update and refine the surrogate model. This iterative process continues until a preset number of cycles is reached, ultimately providing a set of Pareto-optimal antenna layout plans that represent the best trade-offs between the multiple objectives.
Advantages
- Provides multiple Pareto-optimal layout plans in a single run, offering decision-makers and engineers a range of best-compromise solutions between coverage, interference, and cost.
- Significantly improves optimization efficiency by using a surrogate model to approximate expensive objective functions, reducing the number of costly ray-tracing simulations required.
- Enhances planning accuracy by employing a ray-tracing signal propagation model, which is more suitable for simulating 5G high-frequency signals compared to simpler models.
- Automatically optimizes both the number of antennas and their positions, potentially reducing material and installation costs.
- The decomposition-based framework helps maintain solution diversity across the multiple objectives.
- The method is adaptable and can be applied to other communication systems like 3G or 4G.
Applications
- Planning and deployment of 5G Indoor Distributed Antenna Systems (IB-DAS) in various buildings.
- Network optimization for large indoor spaces like airports, railway stations, shopping malls, convention centers, and stadiums.
- Signal coverage design for office buildings, hospitals, universities, and underground facilities (parking lots, subways).
- Integration into 5G network planning software tools specifically tailored for indoor environments.
- Academic and industrial research in the field of wireless network optimization and antenna placement.
