A linear neuromuscular model was developed and incorporated within a driver/vehicle model. Optimal control was used to minimise metabolic energy and path-following error. Simultaneous feed-forward and feedback operation was observed, with the stretch reflex loop acting to reject disturbances. A trade-off between minimising the feedback error signal and energy consumption exists that has not been previously identified. A non-linear, Huxley/Zahalak-based model of an agonist/antagonist muscle pair connected to a second order load was implemented (the ‘MDM’ model). Mechanistic and energy consumption predictions compare favourably with published data. The model was linearized, to allow incorporation within a linear neuromuscular framework. A suitable model structure was fitted using parametric methods. A novel, linear, energy consumption model was proposed. A parameter study of the MDM model was carried out. Variable natural length behaviour was observed, consistent with real muscle operation. Findings suggested that the stretch reflex gain is not large enough to account for low frequency behaviour observed by some researchers for ‘stochastic disturbance’ type experiments. An optimal controller representing cognitive influence was shown to account for this behaviour. A Box-Jenkins method for identifying intrinsic and reflex dynamics models (on the basis of reflex delay) was developed and validated. The impact of the stretch reflex gain and noise levels on identification success was investigated. Intrinsic and reflex models were identified from eight test subjects’ data. The closed-loop neuromuscular model agreed well with measured data, and was generally consistent with MDM model predictions. Low frequency control action and changes in stretch reflex dynamics were observed, stemming from cognitive influence. Other researchers have failed to account for this.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:604261 |
| Date | January 2009 |
| Creators | Hoult, W. |
| Publisher | University of Cambridge |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0017 seconds