Aleatoric Uncertainty from AI-Based 6D Object Pose Predictors for Object-Relative State Estimation
Thomas Jantos, Stephan Weiss, Jan Steinbrener
AI summary
Problem
DL-based 6D object pose predictors lack reliable uncertainty quantification, forcing state estimators to rely on fixed or manually tuned covariance matrices that fail to adapt to inherent prediction noise.
Approach
The method extends any pre-trained pose network with two detached MLPs to predict metric aleatoric uncertainty for translation and rotation, which is then fed directly as dynamic measurement covariance into an Extended Kalman Filter.
Key results
- Extending pre-trained pose predictors with lightweight MLPs for metric aleatoric uncertainty
- Replacing fixed EKF covariance with dynamically predicted uncertainty to eliminate manual tuning
- Enabling uncertainty-driven anchor object switching and outlier rejection
- Demonstrating improved state estimation accuracy on synthetic and real-world datasets
Why it matters
Provides a scalable, low-overhead solution for robust real-time navigation on edge devices by automating measurement trust calibration in probabilistic state estimators.
Abstract
Deep Learning (DL) has become essential in various robotics applications due to excelling at processing raw sensory data to extract task specific information from semantic objects. For example, vision-based object-relative navigation relies on a DL-based 6D object pose predictor to provide the relative pose between the object and the robot as measurements to the robot’s state estimator. Accurately knowing the uncertainty inherent in such Deep Neural Network (DNN) based measurements is essential for probabilistic state estimators subsequently guiding the robot’s tasks. Thus, in this letter, we show that we can extend any existing DL-based object-relative pose predictor for aleatoric uncertainty inference simply by including two multi- layer perceptrons detached from the translational and rotational part of the DL predictor. This allows for efficient training while freezing the existing pre-trained predictor. We then use the inferred 6D pose and its uncertainty as a measurement and cor- responding noise covariance matrix in an extended Kalman filter (EKF). Our approach induces minimal computational overhead such that the state estimator can be deployed on edge devices while benefiting from the dynamically inferred measurement uncertainty. This increases the performance of the object-relative state estimation task compared to a fix-covariance approach. We conduct evaluations on synthetic data and real-world data to underline the benefits of aleatoric uncertainty inference for the object-relative state estimation task.