Adversarial Subgraph Contrastive Learning for Predicting Grasp Stability of Robotic Hands With Multimodal Signals

Accurate prediction of grasp stability is crucial for reliable and precise operations with multi-fingered robotic hands. Traditional methods tend to oversimplify tactile information and pay equal attention to all regions of the data. This can obscure subtle yet critical variations and introduce noise, increasing the risk of stability assessment errors. To address these challenges, a novel self-supervised method, Adversarial Subgraph Contrastive Learning (ASCL), is proposed. It constructs an instance graph from the spatial distribution and features of perceptual nodes. It employs a bi-level adversarial strategy to enhance latent data representations by maximizing the mutual information between the instance graph and its semantic subgraphs, while minimizing it with its noisy subgraphs. To prevent trivial solutions and continuous relaxation of semantic subgraphs, node confidence and edge connection terms are incorporated to ensure stabilization. From an information-theoretic perspective, ASCL exhibits notable advantages on unlabeled or sparsely labeled data, well outperforming existing methods in empirical tests with robotic hands. Note to Practitioners—Grasp stability prediction is vital for the monitor and control of robotic interactions, particularly in industrial and service applications where precision and reliability are important. This paper introduces a self-supervised graph contrastive learning method that constructs a graph structure from tactile and near-contact data in order to capture intricate spatial distributions and feature details. By employing a bi-level adversarial strategy, the method emphasizes meaningful substructures within data, highlighting critical variations while suppressing redundancy and noise. The experimental results demonstrate that significant improvements in grasp stability prediction for downstream tasks involving objects with varying attributes are achieved, effectively reducing the reliance on manual labeling. Furthermore, the model exhibits generalization capabilities to unseen objects. The work represents an advancement in improving robotic fine manipulation and autonomous decision-making skills for intelligent automated systems.