Search papers, labs, and topics across Lattice.
This paper introduces a neuro-symbolic framework designed to enhance multi-pair robotic assembly in unstructured environments by addressing challenges such as spatial interference and contact uncertainties. By hierarchically generating optimal subgraphs for each assembly pair and utilizing a lightweight discriminator to manage edge cases, the framework effectively decouples generality from specific scenarios while maintaining coherence in execution. The approach achieves a remarkable 97.00% global executability in offline evaluations and a 90% success rate in real-robot deployments, significantly surpassing traditional planning methods.
Achieving 97% global executability in robotic assembly, this framework seamlessly integrates cognitive decision-making with physical execution, overcoming traditional limitations.
Multi-pair robotic assembly in unstructured environments faces spatial interference and contact uncertainties. Existing paradigms fail to bridge cognitive decision-making and physical execution, as they either encounter state-space explosion and knowledge bottlenecks or suffer from logical hallucinations and topological conflicts. We propose an end-to-end neuro-symbolic framework that solves the challenge hierarchically: generating optimal subgraphs for each pair, decoupling generality from edge cases, and then resolving cross-pair interferences. Given an eye-on-hand RGB-D assembly scene, the framework extracts semantic instance identity and state while quantifying the scene for divergence calculation. For each pair, optimal subgraph is generated via LLM using barely basic actions to mitigate hallucinations. Supportive actions for edge cases are reasoned and inserted with a lightweight discriminator. Driven by the divergence between the quantified baseline and current scene, it is easily extensible at low cost. Augmented subgraphs are topologically coordinated into global sequences while preserving internal behavioral coherence. Dynamic behavior trees embedding atomic skills close the force-aware execution loop. Offline evaluation on 100 real-world scenes achieves 97.00% global executability, outperforming classical and state-of-the-art planners. Real-robot deployment on a UR3 arm attains 90% success rate with 0.5 mm tolerance under strong interference, demonstrating a unified and verifiable solution for complex autonomous assembly.
