Robots detect and fix assembly failures using smart movement stages
Task aware Dynamic Movement Primitives for failure detection and recovery in contact rich manipulation
Robotics
Summary
Robots often struggle to put parts together when things don't line up exactly right or when they feel unexpected contact forces. The authors created a system that helps robots recognize different steps while assembling, like approaching, aligning, and inserting a peg. The system watches sensor data to spot when something goes wrong and then uses a special search movement to fix the problem before trying again. This way, the robot can recover from small misplacements using just one example demonstration.
What this means in practice
- •For industrial automation teams: Improve robotic assembly lines by enabling robots to detect and recover from contact-related failures automatically during peg-in-hole tasks.
- •For robotic system integrators: Integrate a task-aware failure detection framework to enhance robot adaptability with minimal demonstration and limited sensor inputs.
Authors
Bhavnashri A, Sobia Shafi, Krishnapuram Himavarshini, Anuj Tiwari
Abstract
Assembly remains a challenging robotic manipulation task in presence of tight tolerances and complex contact interactions. While Learning from Demonstration (LfD) frameworks like Dynamic Movement Primitives (DMPs) can effectively encode trajectories from a single demonstration, they are highly sensitive to variations in initial grasp configurations and external contact forces. Such variations often lead to task failures during the contact rich phases. This paper presents a task aware failure detection and recovery framework that integrates DMP based trajectory generation with real time stage classification. Utilizing Quadratic Discriminant Analysis (QDA) trained on multimodal sensor data, the framework segments execution into approach, alignment, and insertion stages for a Peg in Hole (PiH) assembly operation. By using goal relative position data as features, this classification generalizes to unseen goal positions without requiring retraining, matching the inherent generalization capability of DMPs. Anomaly detection is performed online using a Mahalanobis distance metric computed over force features, isolating contact induced failures from nominal trajectory execution. Upon failure detection, a spiral search recovery policy is triggered to actively realign the peg under contact before resuming the learned DMP insertion. The proposed approach is evaluated on an experimental setup achieving 95% stage classification accuracy, and demonstrates reliable failure recovery under lateral misalignments of up to 3 mm using only a single demonstration.