Abstract:Indoor positioning using UWB technology has gained interest due to its centimeter-level accuracy potential. However, multipath effects and non-line-of-sight conditions cause ranging errors between anchors and tags. Existing approaches for mitigating these ranging errors rely on collecting large labeled datasets, making them impractical for real-world deployments. This paper proposes a novel self-supervised deep reinforcement learning approach that does not require labeled ground truth data. A reinforcement learning agent uses the channel impulse response as a state and predicts corrections to minimize the error between corrected and estimated ranges. The agent learns, self-supervised, by iteratively improving corrections that are generated by combining the predictability of trajectories with filtering and smoothening. Experiments on real-world UWB measurements demonstrate comparable performance to state-of-the-art supervised methods, overcoming data dependency and lack of generalizability limitations. This makes self-supervised deep reinforcement learning a promising solution for practical and scalable UWB-ranging error correction.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **Correct the ranging error in UWB (Ultra - Wideband) positioning systems through self - supervised deep reinforcement learning methods without using real - world datasets with labels**. ### Problem Background UWB technology has received extensive attention in indoor positioning due to its centimeter - level positioning accuracy. However, multipath effects and non - line - of - sight (NLOS) conditions can lead to ranging errors between anchor nodes and tags. Existing error - correction methods rely on collecting a large number of labeled datasets, which makes them impractical in actual deployments. Specifically: 1. **Difficulties in data collection**: Collecting data with real - world labels requires specialized equipment and expertise, and is time - consuming and labor - intensive. 2. **Poor generalization ability**: Trained models perform poorly in unseen environments because there are differences in channel impulse response (CIR) and UWB physical - layer characteristics in different environments. ### Paper's Solution To solve the above problems, this paper proposes a self - supervised method based on deep reinforcement learning (RL) that does not require labeled real - world data. Specifically: - **Self - supervised learning framework**: Predict and correct ranging errors by using channel impulse response (CIR) as state input. This method gradually improves the accuracy of error correction by iteratively improving the generated correction values and combining trajectory prediction, filtering, and smoothing techniques. - **Reinforcement learning process**: The RL agent uses CIR as a state and predicts correction values to minimize the error between the corrected and estimated ranges. The agent achieves self - supervised learning by continuously iteratively improving these correction values. - **Experimental verification**: Experiments on real - world UWB measurement data show that the performance of this method is comparable to that of the state - of - the - art supervised methods, overcoming the problems of data - dependence and insufficient generalization ability. ### Main Contributions 1. **First introduction of the self - supervised deep RL framework**: For CIR - based UWB ranging error correction, applicable to two - way ranging (TWR) systems. 2. **Fully self - supervised**: Can be successfully implemented without data collection or relying on real - world labels. 3. **Performance analysis**: Comparative analysis of the proposed self - supervised deep RL framework with the state - of - the - art supervised convolutional neural network (CNN). Through this method, the paper provides a more practical and scalable UWB ranging error correction scheme in practical applications.

Removing the need for ground truth UWB data collection: self-supervised ranging error correction using deep reinforcement learning

Self-Supervised Deep Location and Ranging Error Correction for UWB Localization

An Unsupervised Transfer Learning Method for UWB Ranging Error Mitigation

A Semi-Supervised Learning Approach for Ranging Error Mitigation Based on UWB Waveform

A Probabilistic Learning Approach to UWB Ranging Error Mitigation.

An Attention-Assisted UWB Ranging Error Compensation Algorithm

Self-calibration and Collaborative Localization for UWB Positioning Systems

Deep reinforcement learning for automatic run-time adaptation of UWB PHY radio settings

A Machine Learning Approach to Ranging Error Mitigation for UWB Localization

Ultra-Wideband Ranging Error Mitigation with Novel Channel Impulse Response Feature Parameters and Two-Step Non-Line-of-Sight Identification

Multipath-Assisted Single-Anchor Localization Via Deep Variational Learning

Deep Learning-Based Ranging Error Mitigation Method for UWB Localization System in Greenhouse.

Deep GEM-Based Network for Weakly Supervised UWB Ranging Error Mitigation

Error Mitigation for TDoA UWB Indoor Localization using Unsupervised Machine Learning

Research on Inertial Navigation and Environmental Correction Indoor Ultra-Wideband Ranging and Positioning Methods

A Semi-Supervised Learning Approach for UWB Ranging Error Mitigation

Deep Learning for Ultra-Wideband Indoor Positioning

MR-ULINS: A Tightly-Coupled UWB-LiDAR-Inertial Estimator with Multi-Epoch Outlier Rejection

Indoor Positioning Using Ultrawideband and Inertial Measurements

UWB Indoor Localization Using Deep Learning LSTM Networks

Anchor Self-Localization Algorithm Based on UWB Ranging and Inertial Measurements