Ruprecht-Karls-Universität Heidelberg
Optimization in Robotics and Biomechanics

Modeling and optimal control for mobility assistance and rehabilitation devices

Khai-Long Ho Hoang, Davide Corradi, Katja Mombaur - ORB

This work is part of the MOBOT project (Intelligent Active MObility Assistance RoBOT integrating Multimodal Sensory Processing, Proactive Autonomy and Adaptive Interaction) funded by the European Union within the FP7 program under grant agreement no. 600796.

Our ageing society comes along with an increased number of humans suffering from mobility impairments. Yet, it is important for them to maintain their life quality through independence in daily activities. The MOBOT project gathers the expertise from five academic, two clinical and one industrial partner from five different EU-countries to create a robotic device which provides multimodal mobility assistance to elderly humans. User-centered and context-adaptive natural support is provided proactively by realizing an autonomous and context-specific monitoring of human activities and behavioural patterns as well as adaptively and interactively by performing adaptive compliance control for optimal physical support and fall prevention based on the analysis of multi-sensory and physiological signals related to gait and postural stability.

The tasks realized by the ORB group are strongly tied to our optimal control methods and our expertise in modeling, analysis and synthesis of human motions.

Anthropometric data for elderly people

The biomechanics of human movement can be fundamentally described by dynamic models of the human body. In terms of multi-body systems, it is necessary to know the anthropometric parameters of the body segments. We created equations which provide the geometric and inertial parameters of the major body segments specifically of elderly humans. These parameters can be used in a practicable way in the context of creating a dynamic multi-body model for geriatric motions.

Ho Hoang, K.-L. & Mombaur, K. Adjustments to de Leva-anthropometric regression data for the changes in body proportions in elderly humans Journal of Biomechanics, 2015, 48, 3732 - 3736


Gait patterns and gait performance indicators for elderly people

Robotic mobility assistance devices are supposed to provide a purposeful and adaptive support for elderly humans. To achieve assistance which is specifically adapted to the individual needs and capabilities of the user, the devices should be capable of estimating the overall gait performance as well as the current state of the subject on-line. Signals gathered by on-board sensors can be exploited to obtain, amongst other values, the interaction forces, positions and velocities of the subject. We created methods to identify gait parameters based on easily accessible on-line sensor signals (while explicitly omitting any additional measuring equipment). Based on such signals, we derive temporo-spatial gait performance indicators which can be used by robotic mobility aid platforms to evaluate the performance of the user and apply individualized user-adaptive assistance strategies. By analyzing the correlation between these parameters we propose a set of gait performance indicators and derive performance classes that can be used in combination or isolated.

Ho Hoang, K.-L.; Corradi, D.; Delbasteh, S. & Mombaur, K. Gait performance indicators for elderly humans interacting with robotic mobility assistance devices IEEE International Conference on Rehabilitation Robotics (ICORR), 2015, 758-763


Optimal STS motions and STS support

For elderly people, one of the most challenging type of motions is the sit to stand (STS) transfer. Hence, mobility assistive devices must be able to support STS motions to provide a comprehensive assistance to geriatric subjects. We synthesized the geriatric STS motion and computed the optimal mechanical design parameters of a STS assistance device in two steps.

Firstly, we established a method to predict geriatric STS movements and computed the best possible actions provided by a generic external device to support these movements. We treated three types of hypothetical active devices that act on different parts of the patient’s body and provide different levels of support. Our approach is based on the solution of optimal control problems for a whole-body multi-phase model of humans standing up from sitting to the upright position.

Secondly, in a model-based optimization approach the optimal states and controls as well as the mechanical design parameters of a STS assistance device were computed using direct multiple shooting optimal control methods. The results for the optimal support forces as well as the trajectories oftheir force application points obtained in the first step served as constraints for design optimizations of different device types. Based on the resulting optimal design parameters two different prototypes of the mobility assistance devices weremanufactured which provide STS assistance to two different levels of mobility impairments.

Mombaur, K. & Ho Hoang, K.-L. How to best support sit to stand transfers of geriatric patients: Motion optimization under external forces for the design of physical assistive devices Journal of Biomechanics , 2017, 58, 131 - 138

Ho Hoang, K.-L. & Mombaur, K. Optimal design of a physical assistive device to support sit-to-stand motions IEEE International Conference on Robotics and Automation (ICRA), 2015, 5891-5897


Fall prevention control

To address fall prevention control, we considered a robotic rollator equipped with actuated handles which are capable of motions on the sagittal plane. We developed a methodology, based on optimal control and rigid body models of the user and of the device, to compute optimal handle trajectories to help the user regain an upright, balanced posture starting from perturbed initial conditions while standing without walking.

Corradi, D.; Hoang, K. L. H. & Mombaur, K. Optimal control for balance assistance using a robotic rollator 2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob), 2016, 594-599

K. Mombaur,
Last Update: 12.09.2017 - 14:42

The photographs in the header of this webpage have been taken at the Musee de l'Automate in Souillac, France