Abstract:Purpose Assembly action recognition plays an important role in assembly process monitoring and human-robot collaborative assembly. Previous works overlook the interaction relationship between hands and operated objects and lack the modeling of subtle hand motions, which leads to a decline in accuracy for fine-grained action recognition. This paper aims to model the hand-object interactions and hand movements to realize high-accuracy assembly action recognition. Design/methodology/approach In this paper, a novel multi-stream hand-object interaction network (MHOINet) is proposed for assembly action recognition. To learn the hand-object interaction relationship in assembly sequence, an interaction modeling network (IMN) comprising both geometric and visual modeling is exploited in the interaction stream. The former captures the spatial location relation of hand and interacted parts/tools according to their detected bounding boxes, and the latter focuses on mining the visual context of hand and object at pixel level through a position attention model. To model the hand movements, a temporal enhancement module (TEM) with multiple convolution kernels is developed in the hand stream, which captures the temporal dependences of hand sequences in short and long ranges. Finally, assembly action prediction is accomplished by merging the outputs of different streams through a weighted score-level fusion. A robotic arm component assembly dataset is created to evaluate the effectiveness of the proposed method. Findings The method can achieve the recognition accuracy of 97.31% and 95.32% for coarse and fine assembly actions, which outperforms other comparative methods. Experiments on human-robot collaboration prove that our method can be applied to industrial production. Originality/value The author proposes a novel framework for assembly action recognition, which simultaneously leverages the features of hands, objects and hand-object interactions. The TEM enhances the representation of dynamics of hands and facilitates the recognition of assembly actions with various time spans. The IMN learns the semantic information from hand-object interactions, which is significant for distinguishing fine assembly actions.

Repetitive assembly action recognition based on object detection and pose estimation

Learning-Based Hand Motion Capture and Understanding in Assembly Process

A Skeleton-Based Assembly Action Recognition Method with Feature Fusion for Human-Robot Collaborative Assembly

A Novel Multi-Stream Hand-Object Interaction Network for Assembly Action Recognition

Hybrid Convolutional Neural Network Approaches for Recognizing Collaborative Actions in Human–Robot Assembly Tasks

Monitoring of Assembly Process Using Deep Learning Technology.

Online Robust Action Recognition Based on a Hierarchical Model

Deep learning based assembly process action recognition and progress prediction facing human-centric intelligent manufacturing

Human Action Recognition Based on Kinematic Similarity in Real Time.

Human-object integrated assembly intention recognition for context-aware human-robot collaborative assembly

A posture detection method for augmented reality–aided assembly based on YOLO-6D

Human Action Recognition Using Deep Learning Methods.

How Object Information Improves Skeleton-based Human Action Recognition in Assembly Tasks

A digital twin-driven human-machine interactive assembly method based on lightweight multi-target detection and assembly feature generation

Multi-class Assembly Parts Recognition Using Composite Feature and Random Forest for Robot Programming by Demonstration.

Data-efficient Multimodal Human Action Recognition for Proactive Human–robot Collaborative Assembly: A Cross-Domain Few-Shot Learning Approach

Algorithm of Human Posture Action Recognition and Imitation for Robots

Online Action Detection and Forecast Via Multitask Deep Recurrent Neural Networks

Development of an action classification method for construction sites combining pose assessment and object proximity evaluation

6D Assembly Pose Estimation by Point Cloud Registration for Robot Manipulation

Human Action Recognition and Counting Based on Embedded System