Context-Aware Human Activity Recognition in Industrial Processes

Abstract

The automatic, sensor-based assessment of human activities is highly relevant for production and logistics, to optimise the economics and ergonomics of these processes. One challenge for accurate activity recognition in these domains is the context-dependence of activities: Similar movements can correspond to different activities, depending on, e.g., the object handled or the location of the subject. In this paper, we propose to explicitly make use of such context information in an activity recognition model. Our first contribution is a publicly available, semantically annotated motion capturing dataset of subjects performing order picking and packaging activities, where context information is recorded explicitly. The second contribution is an activity recognition model that integrates movement data and context information. We empirically show that by using context information, activity recognition performance increases substantially. Additionally, we analyse which of the pieces of context information is most relevant for activity recognition. The insights provided by this paper can help others to design appropriate sensor set-ups in real warehouses for time management.

EventTube: An Artificial Intelligent Edge Computing Based Event Aware System to Collaborate With Individual Devices in Logistics Systems

“…A. Experiment Settings 1) Data sources and analysis: our dataset consists of video data captured on RGB-Monoculars from mobile robots in logistics warehouses, logistic activity recognition dataset (LARa) [18] and public autonomous driving dataset KITTI [19]. The logistics warehouse site covers various obstacle and movement scheduling scenarios, and the video quality varies depending on the movement speed and lighting conditions when the RGB-Monoculars are capturing the data.…”

Skeleton-based human action recognition with context awareness

“…In this paper, we propose a novel human action recognition framework with context awareness, which utilizes both the temporal evolution of the skeleton data and the contextual information of the scene to achieve better action recognition performance.The major contributions of this work lie in two aspects: (1) We propose a scheme combined with YoloV5 and ST-GCN for human action recognition taking the contextual information of the scene into consideration. (2) We propose new joint features in designing graph convolutional neural network which can use contextual information.…”

“…However, the existing skeleton-based action recognition methods mainly focus on modeling the temporal evolution of the skeleton data, while not taking the contextual information of the scene into consideration. The context information of the scene, such as the environment and the interaction between the actors, can provide additional information that is essential for action recognition especially in industrial scenario [1] as figure.1 shown.…”