SWiLoc: Fusing Smartphone Sensors and WiFi CSI for Accurate Indoor Localization

Khairul Mottakin, Kiran Davuluri, Mark Allison, Zheng Song

TL;DR

This paper introduces SWiLoc, a novel system that fuses smartphone sensor data with passive WiFi Channel State Information (CSI) to achieve highly accurate indoor localization. By introducing "Correction Zones" and a two-phase approach, SWiLoc significantly reduces localization errors compared to existing methods, even when users hold their phones in arbitrary orientations.

What Problem Does This Paper Address?

Accurate indoor localization is crucial for many applications, but GPS is unreliable indoors. While various indoor localization techniques exist, they often suffer from limitations. Pedestrian Dead Reckoning (PDR), which uses smartphone sensors, is promising but struggles with inaccurate walking direction estimation due to misalignment between the phone's orientation and the user's actual movement. Existing methods to improve this either require external landmarks, extensive user-specific training, or are sensitive to how the phone is held. This paper aims to overcome these limitations by developing a system that provides accurate and robust indoor localization without requiring user training or specific phone-holding postures.

Methods

SWiLoc employs a two-phase approach:

  • Phase 1: Correction Zone Sensing: When a user enters a "Correction Zone" (equipped with WiFi routers and receivers), SWiLoc uses passive WiFi CSI data along with smartphone motion sensor data (step count, pitch, roll, azimuth) to determine the user's walking direction. Crucially, it establishes a correlation between the phone's orientation and the user's actual movement direction within this zone.
  • Phase 2: Direction and Location Estimation: After leaving a correction zone, SWiLoc uses the established correlation from Phase 1 to infer the user's walking direction based solely on the phone's orientation. This corrected direction is then used to estimate the user's precise location.

Key innovations include:

  • Correction Zones: Strategically placed WiFi access points and receivers create zones where WiFi CSI can be reliably captured.
  • Fresnel Zone-Based Approach: SWiLoc utilizes a Fresnel zone-based method for WiFi sensing, which is geometry-based and doesn't require training. To address issues with unreliable directions (parallel to LoS), SWiLoc uses four receivers instead of the traditional two.
  • Geometric Relationship: Instead of relying on ratios, SWiLoc uses a geometric relationship between movement and Fresnel zones for more accurate direction estimation.
  • Phone Orientation Mapping: A novel mapping technique is used to translate the phone's orientation into the user's actual walking direction, making the system robust to how the phone is held.

Datasets

The paper does not explicitly mention training datasets. Evaluation was performed using data collected from experiments in two real-world indoor environments: an office room and a long corridor. These environments were instrumented with the SWiLoc system.

Key Insights and Ideas

  • Fusing WiFi CSI and PDR is powerful: Combining passive WiFi sensing with active smartphone sensors (PDR) offers a synergistic approach that overcomes the individual limitations of each.
  • Correction Zones are key enablers: These zones act as calibration points, allowing the system to learn and adapt to the user's movement and phone orientation.
  • Geometry over Ratios: Using a geometric understanding of Fresnel zones, rather than simple ratios, leads to more accurate direction estimation, especially in challenging scenarios.
  • Phone Holding Posture is a solvable problem: The mapping technique effectively decouples the user's direction from the phone's orientation, making the system robust to variations in how the phone is carried.
  • Accuracy gains are substantial: The proposed system significantly outperforms existing methods in both direction and location estimation.

Key Results

  • Direction Estimation Accuracy (Phase 1): Achieved an average 75th percentile error of 8.89 degrees.
  • Location Estimation Accuracy (Overall): Achieved an 80th percentile error of 1.12 meters, representing a 49% reduction compared to state-of-the-art methods.
  • Comparison with State-of-the-Art: SWiLoc demonstrated superior performance against methods like WiDir, Widar, and WalkCompass, achieving a 64% reduction in 75th percentile error compared to the closest competitor.
  • Robustness to Phone Holding: The system maintained high accuracy regardless of whether the phone was held in hand, shirt pocket, or pant pocket.

Limitations and Future Work

  • Multi-person Interference: The system's accuracy can decrease when multiple individuals are in the same correction zone simultaneously.
  • Minimum Post-LoS Distance: Users need to walk at least 0.5 meters after crossing a Line-of-Sight (LoS) to ensure effective correction.
  • Phone Orientation Assumption: Phase 2 relies on the assumption that the mapping between user direction and phone orientation remains constant. If a user significantly changes how they hold their phone, multiple correction zones are needed to re-establish this mapping.
  • Future Work: The authors plan to integrate SWiLoc with other sensor fusion-based approaches (like WalkCompass) and explore optimal placement and number of correction zones for comprehensive coverage. They also aim to address the multi-person interference issue.

Summary updated

Abstract

Dead reckoning is a promising yet often overlooked smartphone-based indoor localization technology that relies on phone-mounted sensors for counting steps and estimating walking directions, without the need for extensive sensor or landmark deployment. However, misalignment between the phone’s direction and the user’s actual movement direction can lead to unreliable direction estimates and inaccurate location tracking. To address this issue, this paper introduces SWiLoc, an enhanced direction correction system that integrates passive WiFi sensing with smartphone-based sensing to form Correction Zones. Our two-phase approach accurately measures the user’s walking directions when passing through a Correction Zone and further refines successive direction estimates outside the zones, enabling continuous and reliable tracking. In addition to direction correction, SWiLoc extends its capabilities by incorporating a localization technique that leverages corrected directions to achieve precise user localization. This extension significantly enhances the system’s applicability for high-accuracy localization tasks. Additionally, our innovative Fresnel zone-based approach, which utilizes unique hardware configurations and a fundamental geometric model, ensures accurate and robust direction estimation, even in scenarios with unreliable walking directions. We evaluate SWiLoc across two real-world environments, assessing its performance under varying conditions such as environmental changes, phone orientations, walking directions, and distances. Our comprehensive experiments demonstrate that SWiLoc achieves an average 75th percentile error of 8.89 degrees in walking direction estimation and an 80th percentile error of 1.12 meters in location estimation. These figures represent reductions of 64% and 49%, respectively for direction and location estimation error, over existing state-of-the-art approaches.
BibTeX
@misc{mottakin2024swiloc,
  author = {Khairul Mottakin and Kiran Davuluri and Mark Allison and Zheng Song},
  title = {{SWiLoc: Fusing Smartphone Sensors and WiFi CSI for Accurate Indoor Localization}},
  year = {2024},
  doi = {10.3390/s24196327},
}
Figure