RadProPoser: A Framework for Human Pose Estimation with Uncertainty Quantification from Raw Radar Data

2025-08-05

AI Summary

Radar-based human pose estimation (HPE) offers inherent privacy preservation and illumination invariance but suffers from noise corruption and multipath interference. To address these challenges, we propose the first end-to-end uncertainty-aware framework for radar tensor-to-3D pose estimation. Our method operates directly on complex-valued radar inputs and integrates variational inference with a probabilistic encoder-decoder architecture, enabling heteroscedastic noise modeling and total uncertainty calibration. We systematically compare Gaussian and Laplacian priors and likelihoods, and introduce latent-variable sampling for effective data augmentation. Evaluated on a custom optical motion-capture dataset, our approach achieves a mean joint error of 6.43 cm (5.68 cm at 45° viewing angle), an uncertainty calibration error of only 0.021, and an F1-score of 0.870 on downstream tasks—demonstrating substantial improvements in reliability, robustness, and interpretability of radar HPE.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Radar-based human pose estimation (HPE) provides a privacy-preserving, illumination-invariant sensing modality but is challenged by noisy, multipath-affected measurements. We introduce RadProPoser, a probabilistic encoder-decoder architecture that processes complex-valued radar tensors from a compact 3-transmitter, 4-receiver MIMO radar. By incorporating variational inference into keypoint regression, RadProPoser jointly predicts 26 three-dimensional joint locations alongside heteroscedastic aleatoric uncertainties and can be recalibrated to predict total uncertainty. We explore different probabilistic formulations using both Gaussian and Laplace distributions for latent priors and likelihoods. On our newly released dataset with optical motion-capture ground truth, RadProPoser achieves an overall mean per-joint position error (MPJPE) of 6.425 cm, with 5.678 cm at the 45 degree aspect angle. The learned uncertainties exhibit strong alignment with actual pose errors and can be calibrated to produce reliable prediction intervals, with our best configuration achieving an expected calibration error of 0.021. As an additional demonstration, sampling from these latent distributions enables effective data augmentation for downstream activity classification, resulting in an F1 score of 0.870. To our knowledge, this is the first end-to-end radar tensor-based HPE system to explicitly model and quantify per-joint uncertainty from raw radar tensor data, establishing a foundation for explainable and reliable human motion analysis in radar applications.

Problem

• Estimating human pose from noisy radar data
• Quantifying uncertainty in joint position predictions
• Enabling privacy-preserving human motion analysis

Innovation

• Probabilistic encoder-decoder for radar pose estimation
• Variational inference for joint keypoint uncertainty prediction
• Calibratable uncertainty modeling with Gaussian and Laplace distributions