Computational Modeling and Control for Personalized Neuroprosthetics and Rehabilitation
2015
Human motor function consists of different levels of components from peripheral to central mechanism. This HDR thesis focuses on the computational modeling aspects, however it is oriented to contribute to modeling and control of these different levels from peripheral to central: From local neuroprosthetic muscle control to central motor learning control.
Our INRIA Demar team is specialized first for neuroprosthetics and functional electrical stimulation (FES). I have been in charge of modeling aspects for personalized identification of stimulus-evoked neuromuscular dynamics as in Chap.2. It is extended into personalized neuroprosthetics as in Chap.3 to predict subject-specific torque estimation by means of evoked electromyography (eEMG) even under muscle fatigue, which is fundamental problem over FES. The EMG Feedback Predictive Muscle Control in FES is archieved including real-time performance together with wireless portable stimulator as in Chap.4. Physiological muscle modeling under FES as in Chap.2 is extended also for volitional motor actions. The work of inversed muscle activation solution using muscle synergy extraction is also proposed in Chap.5 envisaging for finding redundant muscle activation patterns in multi-channel neuroprosthetics. Chap.6 is regarding personalized home rehabilitation, especially focusing on the development of personalized balance measure considering subjectspecific differences by adaptive identification. The rehabilitation in general is indeed a re-learning process under the given dynamics constraints. Chap.7 deals with developing a novel computational motor control/learning paradigm to understand how redundancy is managed in human central nervous system.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI