WhiteBox

Funding

LOEWE-Schwerpunktantrag, granted by the Hessisches Ministerium für Wissenschaft und Kunst from 1. 2021 to 12.2024

Consortium

Nine academic partners from different departments of TU Darmstadt are involved in this project. You can see details here,

The LOEWE Research Cluster WhiteBox aims to develop methods at the intersection between Cognitive Science and AI to make human and artificial intelligence more understandable. WhiteBox is a joint project with nine research partners from different departments of TU Darmstadt, from engineering and computer science to human science. Lauflabor group Leads WP3.1 “Man and Machine walk together” in this project which started in 2021. Here we develop the bioinspired human gait models and connect them to robot models. Our research focuses on developing Whitebox biomechanical models of human locomotion and motor control to be integrated with the black-box (learning-based) models for robot control to improve our understanding of locomotion control in humans and robots. The outcomes will be later used to develop bioinspired robots and assistive devices.

Objectives

How can one gain an improved understanding of human walking and understand the systematic adaptation of gait patterns to contexts such as environmental situations? How can movement models be obtained that enable the development of new diagnostics and technical approaches for movement support (e.g. for rehabilitation or training)?

Further Info

Human walking behavior is not sufficiently understood from a biomechanical or behavioral point of view. Prescriptive models can precisely predict gait behavior based on data, while descriptive models (based on biomechanics) can in principle explain gait behavior, but have not yet allowed quantitatively accurate predictions. We aim to combine the positive properties of both approaches by using the approximation capability of the Deep Neural Networks (DNN) as a black box model, combining them with unchangeable physical principles and thus changing the direction of the white box.

Objectives

How can one gain an improved understanding of human walking and understand the systematic adaptation of gait patterns to contexts such as environmental situations? How can movement models be obtained that enable the development of new diagnostics and technical approaches for movement support (e.g. for rehabilitation or training)?

Further Info

Human walking behavior is not sufficiently understood from a biomechanical or behavioral point of view. Prescriptive models can precisely predict gait behavior based on data, while descriptive models (based on biomechanics) can in principle explain gait behavior, but have not yet allowed quantitatively accurate predictions. We aim to combine the positive properties of both approaches by using the approximation capability of the Deep Neural Networks (DNN) as a black box model, combining them with unchangeable physical principles and thus changing the direction of the white box.