New QAACF filter boosts foot-mounted indoor navigation accuracy
Researchers slash computational cost while improving dead reckoning precision.
Pedestrian Dead Reckoning (PDR) is critical for indoor navigation where GPS suffers from signal degradation due to roofs and high-rise buildings. A team of researchers introduces the Quaternion-Averaging-Based Adaptive Complementary Filter (QAACF) to improve the accuracy of foot-mounted Attitude and Heading Reference Systems (AHRS) while cutting computational overhead. The filter fuses angular velocity-derived quaternions with those from acceleration and magnetic field measurements using Markley's quaternion averaging, a more rigorous combination method than linear interpolation. QAACF adaptively adjusts the weights of each sensor input according to gait phases and magnetic disturbance levels.
Experimental results show QAACF achieves lower Root Mean Square Errors (RMSE) compared to existing attitude estimation filters, all while requiring significantly less computational cost than Kalman filters. This makes it particularly appealing for real-time applications on resource-constrained wearable devices. The work, published on arXiv (cs.RO) in July 2026, promises to enhance indoor navigation systems in smartphones, AR/VR headsets, and robotics, providing reliable position estimation without external infrastructure.
- QAACF uses Markley's quaternion averaging for more accurate sensor fusion than linear interpolation.
- Adaptive weight adjustment based on gait phases and magnetic disturbances improves robustness.
- Achieves lower RMSE than existing filters with lower computational cost than Kalman filters.
Why It Matters
Enables accurate, low-cost indoor navigation for wearables and robots without GPS dependency.