Conformal Prediction for Indoor Positioning with Correctness Coverage Guarantees

TL;DR

The paper addresses the unreliable uncertainty inherent in deep learning–based indoor positioning by introducing conformal prediction (CP) to achieve statistical guarantees on position estimates. It adopts Split CP with non‑conformity scores, including a distance‑based score to reflect spatial errors, and extends the framework with conformal risk control for false discovery and false negative rates in path navigation, plus a conformal p‑value framework for reliability assessment. Experiments on the UJIIndoLoc dataset using lightweight models (MobileNetV1, VGG19, MobileNetV2, ResNet50, EfficientNet) show that CP closely matches target coverage across models, with EfficientNet delivering smaller prediction sets and thus higher certainty. The work provides practical, model‑agnostic reliability guarantees for indoor positioning, enabling robust, real‑time deployments and offering directions for future work in more dynamic environments and richer sensor configurations.

Abstract

With the advancement of Internet of Things (IoT) technologies, high-precision indoor positioning has become essential for Location-Based Services (LBS) in complex indoor environments. Fingerprint-based localization is popular, but traditional algorithms and deep learning-based methods face challenges such as poor generalization, overfitting, and lack of interpretability. This paper applies conformal prediction (CP) to deep learning-based indoor positioning. CP transforms the uncertainty of the model into a non-conformity score, constructs prediction sets to ensure correctness coverage, and provides statistical guarantees. We also introduce conformal risk control for path navigation tasks to manage the false discovery rate (FDR) and the false negative rate (FNR).The model achieved an accuracy of approximately 100% on the training dataset and 85% on the testing dataset, effectively demonstrating its performance and generalization capability. Furthermore, we also develop a conformal p-value framework to control the proportion of position-error points. Experiments on the UJIIndoLoc dataset using lightweight models such as MobileNetV1, VGG19, MobileNetV2, ResNet50, and EfficientNet show that the conformal prediction technique can effectively approximate the target coverage, and different models have different performance in terms of prediction set size and uncertainty quantification.

Figures (6)

• Figure 1: RSSI feature classification.
• Figure 2: Path positioning.
• Figure 3: Training and testing loss.
• Figure 4: Training and testing accuracy.
• Figure 5: Empirical and target coverage.

Theorems & Definitions (2)

• proof
• proof : Proof