An Integration of Deep Neural Network-Based Extended Kalman Filter (DNN-EKF) Method in Ultra-Wideband (UWB)...

Area of Science:

• Robotics
• Artificial Intelligence
• Signal Processing

Background:

• Accurate indoor positioning is crucial for robot navigation in confined environments like homes and warehouses.
• Existing localization methods struggle with precision and reliability in dynamic or noisy indoor settings.
• Ultra-Wideband (UWB) technology offers potential for high-accuracy ranging but requires sophisticated processing.

Purpose of the Study:

• To develop a highly accurate and robust indoor localization algorithm for mobile robots.
• To enhance robot navigation capabilities in challenging, confined indoor spaces.
• To integrate Deep Neural Networks (DNNs) with the Extended Kalman Filter (EKF) for superior localization performance.

Main Methods:

• Developed a novel algorithm, DNN-EKF, combining Deep Neural Networks (DNNs) and the Extended Kalman Filter (EKF).
• Utilized Ultra-Wideband (UWB) technology for precise ranging measurements.
• Evaluated the DNN-EKF method against traditional and existing NN-EKF and LPF-EKF approaches.

Main Results:

• The DNN-EKF algorithm demonstrated superior indoor localization accuracy compared to NN-EKF, LPF-EKF, and traditional methods.
• Achieved the least distance loss, indicating optimal performance in localization precision.
• The proposed method proved effective even in dynamic and noisy indoor environments.

Conclusions:

• The DNN-EKF approach provides a significant advancement in indoor robot localization accuracy and reliability.
• This method is highly suitable for real-time robotic applications requiring precise navigation.
• Integration of DNNs with EKF and UWB technology offers a robust solution for complex indoor navigation challenges.