|M.Sc Thesis||Department of Electrical Engineering|
|Supervisor:||Prof. Emeritus Inbar Gideon|
|Full Thesis text|
The human nervous system controls hundreds of muscles and dozens of joints to create posture and smooth movements. The human motor control is a modular hierarchical system that includes various levels of control. The main levels are the spinal cord, the cerebellum, midbrain ganglions and some specific areas in the cerebral cortex.
Bimanual movement is a movement that involves both arms for an execution of the specified task. Examples of a bimanual movement produced by humans during their routine activity are various: moving of heavy objects by involving two arms, car driving etc. Observations show that involvement of a second arm in a task execution plays important role in improvement of the execution performance, especially in cases of unknown environmental dynamics, unknown mass of a load, unknown force field, random disturbances etc.
Results of an experiment, that was produced with a subject having a motor control problems (essential tremor), have shown significant improvement of movement performance during the production of a planned movement by the two hands together. The two manipulators model have demonstrated an improvement of performance relative to the single manipulator when sinusoidal noises were added to the joints’ torques.
This work analyses the mathematical model of the control of the bimanual human movement. The model of the two arms interconnected by a rigid handle contains the kinematics and dynamics of the planar direct driven manipulators with two degrees of freedom (2DOF). The internal model of the inverse dynamics is modeled by an adaptive controller. Feedback error learning (FEL) algorithm for systems with a feedback delay is used as an algorithm for controller adaptation. Signals delay typical to a nervous system is introduced by a feedback delay in the control scheme.
Organizing of control elements in a task-specified unit is modeled by an adaptive controller common for both arms. Parameters adaptation of the internal model describes a “slow” feedback path of the motor control system, i.e. control path that involves spinal cord and Cerebral Cortex. Fast feedback (spinal cord level) is modeled by a PD controller, producing feedback torques proportional to the trajectory error.