A Constrained Kalman Filter for Wi-Fi-Based Indoor Localization with Flexible Space Organization

Area of Science:

• Robotics and Automation
• Wireless Communication Systems
• Indoor Positioning Technologies

Background:

• Accurate indoor localization is crucial for logistics and mobile device applications.
• Existing Wi-Fi-based localization methods often struggle with dynamic environments and object movement.
• Kalman filters are widely used for state estimation but can be computationally intensive and require model tuning.

Purpose of the Study:

• To develop a constrained Kalman filter for enhanced Wi-Fi-based indoor localization.
• To incorporate object speed constraints for improved accuracy and robustness.
• To provide a numerically optimized filter for efficient real-time computation.

Main Methods:

• Implementation of a constrained Kalman filter algorithm.
• Inclusion of object speed and motion constraints within the filter's state estimation.
• Numerical optimization of the filter for computational efficiency.
• Experimental validation using a robot in a large-scale warehouse environment.

Main Results:

• The constrained Kalman filter demonstrated a 19% improvement in localization accuracy compared to standard methods.
• The optimized filter provided fast computation suitable for real-time applications.
• The approach proved effective in a complex, large-scale (6000 m²) warehouse setting.

Conclusions:

• Constrained Kalman filtering offers a significant advancement in Wi-Fi-based indoor localization accuracy.
• The proposed method is well-suited for dynamic environments and applications requiring precise positioning.
• Numerical optimization enables efficient implementation in practical, real-world scenarios.