Use this URL to cite or link to this record in EThOS:
Title: Mechanisms of human arm motion planning in the presence of multiple solutions
Author: Kodl, Jindrich
ISNI:       0000 0004 2696 3630
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2011
Availability of Full Text:
Access from EThOS:
Access from Institution:
How do humans choose their movement and what are the mechanisms involved in motion planning? This thesis explores the interaction of the central nervous system (CNS) and the external environment, focusing on mechanism it employs for successful execution of fast, complex hand movements. Previous studies have generally assumed that the motion results from the optimisation of a cost function with single optimum. However, this task, similar to many day-to-day tasks, can be performed by more than just one solution. An environment was created that allowed for investigation of complex fast movements. Initially the subjects had to navigate through number of target confi gurations and their respective orientations. The results indicated that the subjects generally utilise multiple plans to achieve the same task. Further experiments presented subjects with alternative trajectories. The results show that the memory of previous motor exploration influences the choice of particular trajectory for explored and unexplored orientations, providing evidence for a motor plan. Analysis of solutions in diff erent directions shows that the choice of a plan depends on previous experience as well as characteristics of motion execution. This choice can be modelled as a Markov process that describes CNS' selection process and how exploration affects it. Considering the results, a computational model was developed, incorporating a set of patterns, which allow generation of successful movements despite large motor variability. A sequence of patterns, a plan, is first prepared and when the movement is executed, accuracy is realised by online prediction of the motion through forward model that utilises derived families of strokes for each pattern. The model takes the visual feedback and by interpolating the corresponding pattern strokes onto the completed trajectory predicts the future trajectory, applying corrective movements if necessary. Despite the feedback delay the outputs demonstrate successful recreation of the observed experimental results.
Supervisor: Burdet, Etienne Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral