- Susanne Blex, Tim Sziburis, Ioannis Iossifidis
- Department of Computer Science, Ruhr West University of Applied Sciences, 45407 Mülheim an der Ruhr, Germany
- Institute of Neuroinformatics, Ruhr University Bochum, 44801 Bochum, Germany
In our research, we model human upper-limb motion by means of the attractor dynamics approach as a promising candidate for the generation of human-like trajectories. For this purpose, we introduce a systematic dataset of 3D center-out hand movements measured by an Intertial Measurement Unit (IMU) attached to a cylindric transport object. Former studies in the context of upper-limb motion were mainly conducted under rather restricted settings, such as focusing on 2D configurations. The recording and analysis of 3D movements in a natural task setting hold additional challenges but are essential for a profound understanding of human motion and our planned mathematical modeling.
In contrast to IMUs, infrared or optical sensors provide a higher precision for motion recordings. This advantage comes at the expense of reduced embedded applicability. An integration in embedded systems needs to fulfil non-functional requirements regarding physical dimensions and energy management. Networks of multiple IMUs are not suitable when solely focusing on hand transport trajectories. Within the VAFES project (Virtual-Reality-based Machine Learning for Arm-Hand Function Evaluation and Support System), we develop an easy-to-use diagnostic glove for movement disorders. This may be applied under clinical as well as ambulant conditions requiring a high extent of portability and a fast setup.
The data are gathered from both hands of young adults without known movement disorders or impairments. The task consists of the transportation of a small cylinder whose movement is recorded by a single IMU of a state-of-the-art motion capture system. The measurement procedure is automatized; for each trial, one of nine targets aligned on a semicircle is randomly chosen. To avoid rhythmic movement patterns and specific time dependencies, random delays are introduced between a visual target cue and an acoustic start signal.
Besides a general analysis of motion trajectories, the experiments will provide insights regarding the variability between trials (intra-subject) and participants (inter-subject), respectively. In the future, we will additionally capture motion trajectories of patients suffering from movement disorders like Parkinson’s disease. We aim to model and compare these trajectories and to find deviations which may allow pathological analysis and diagnosis. Furthermore, these data can be utilized to extend the recent research on the relationship between variability and pathology.