LOPAL: Local Performance-Aware Active Learning from Imperfect Demonstrations
2026-06-15 • Robotics
Robotics
AI summaryⓘ
The authors developed a method called LOPAL to help robots learn tasks better from human demonstrations, even when the demonstrations are sometimes imperfect. Their method uses a model to understand which parts of the demonstration are high quality and then improves on those by creating better robot movements. Additionally, the system asks humans for extra help only when needed to fix weak spots, making learning more efficient. They tested this on a pipe inspection task and found it improved robot performance by over 27% while reducing how much effort the human needs to give.
Learning from Demonstration (LfD)Active LearningGaussian Mixture Model (GMM)Trajectory EncodingLocal Performance AssessmentShared AutonomyRobot Skill AcquisitionData CollectionSimulationTask Performance Improvement
Authors
Johannes Heidersberger, Shail Jadav, Dongheui Lee
Abstract
Learning from Demonstration (LfD) enables intuitive robot skill acquisition by allowing robots to learn directly from human task demonstrations. However, current methods often fail to address the fact that due to suboptimal and inconsistent human behavior, the quality of the demonstration can vary within each demonstration. Therefore, we introduce LOPAL (LOcal Performance-aware Active Learning), an active learning approach that leverages this local demonstration quality information. Our approach consists of two synergistic components. First, a local performance-driven LfD method uses a Gaussian Mixture Model (GMM) to encode both the demonstrated trajectories and their associated local quality assessments. This enables the generation of trajectories that outperform the imperfect demonstrations by utilizing complementary local data of high performance. Second, active data acquisition allows to improve beyond the imperfect demonstrations by collecting additional informative samples. In areas missing good data, the user is actively requested to provide corrections through a shared autonomy (SA) mechanism, while the robot autonomously executes the learned behavior. The efficacy of LOPAL was validated in both a simulation and a real-world experiment. The results from a real-world pipe inspection task showed that the proposed approach can achieve up to 27.31 % improvement in task performance while also reducing the effort required to collect the demonstrations.