Robot Assisted Training to Support Motor Learning for Different Skill Levels

Ekin Basalp, ETH Zürich


When attempting to learn a new motor task, humans require extra source of information about their performance. Generally, availability of external information accelerates the motor learning and increases the quality of the executed movement, which is in line with the underlying principle of Challenge Point Framework.

Challenge Point Framework states that i) learning is dependent on the available and interpretable information during performance and ii) presence of information acts as a challenge to the learner. Thus, learning can be best accomplished if the optimal amount of information is provided, i.e. at the optimal challenge point. To benefit from challenge the learner has to overcome it, which is dependent on his/her skill level.

Since different motor tasks incorporate different types and amount of information, the inherent challenge should be adjusted adequately to the beginners and experts. When suitable types of external information can be supplied at a specific learning phase and updated to more appropriate types in later phases, the entire learning process can be supported.

However, it is yet unknown how the optimal challenge point can be designated and mediated for different skill levels, especially for complex motor tasks. Simulators offer many possibilities to modulate the information available for the trainee in terms of rendering task characteristics and training conditions.

In this project, we investigate the effect of modulating the difficulty of training conditions in our rowing simulator. The motor skill to be learned is upper body-arm sweep rowing. By holding an actual oar handle, subjects try to follow a previously recorded trajectory drawn on the screen. The simulator is then used to increase the virtual water density that will result in noticeable force changes on the handle. Starting from the lowest density condition, in which the task execution effort is minimum, water density will be increased to increment water resistance and effort.

Experimental procedure takes place on three consecutive days. Recorded data are analyzed to find progress of participants’ short- and long-term learning. At the last day, three transfer tests are also assessed to generalize the outcomes of the robot-assisted training in varying training conditions