Difference between revisions of "Sean G. Carver's Research Interests"
(→System Identification of Cooperative Control) |
(→Control Theory Where Agent(s) Rely on Noisy Clocks) |
||
| Line 12: | Line 12: | ||
== Control Theory Where Agent(s) Rely on Noisy Clocks == | == Control Theory Where Agent(s) Rely on Noisy Clocks == | ||
| + | |||
| + | No clock tells time perfectly. In the past, when timekeeping was not good enough for a particular engineering application, the engineers invariably built better clocks. They did not create algorithms to deal with noisy clocks. Thus, up to to now, it has been safe to assume perfection in a controller's timekeeping. Unfortunately, evolution has different constraints, and it is apparent an extension of control theory is needed to deal with the case that clocks are noisy. To date, surprisingly little has been done in this regard. | ||
== Pedagogy for Training Students on System Identification == | == Pedagogy for Training Students on System Identification == | ||
I am interested in finding effective methods for training students in the art of system identification. Training students to use SysID is easier than might first be thought, because one can collect the experimental data needed from ''simulations''. For training and testing purposes, using a simulated system to generate the data is desired, because the experimenter then knows the correct model ahead of time. Knowing the correct model allows an assessment of the suitability and effectiveness of the SysID methods. Moreover, an instructor can create a dataset from simulation, but withhold from students the structure and parameterization of the generating model, giving students the more realistic experience of not already knowing the answer. In my mind, the primary reason this feature is such an advantage appears comes from the fact that no large investment in a laboratory equipment and training is needed. | I am interested in finding effective methods for training students in the art of system identification. Training students to use SysID is easier than might first be thought, because one can collect the experimental data needed from ''simulations''. For training and testing purposes, using a simulated system to generate the data is desired, because the experimenter then knows the correct model ahead of time. Knowing the correct model allows an assessment of the suitability and effectiveness of the SysID methods. Moreover, an instructor can create a dataset from simulation, but withhold from students the structure and parameterization of the generating model, giving students the more realistic experience of not already knowing the answer. In my mind, the primary reason this feature is such an advantage appears comes from the fact that no large investment in a laboratory equipment and training is needed. | ||
Revision as of 20:23, 25 May 2013
Broadly, my interests involve system identification (SysID) applied to biological systems, especially:
- developing methods of SyID,
- applying methods of SysID to understand biological processes, and
- advancing methods for teaching SysID (pedagogy).
By system identification, I mean using statistics, derived from experimental data, to constrain models of the system.
System Identification of Cooperative Control
When people cooperate to perform a task that requires that they stay in sync (e.g. salsa dancing, or chamber music), how do they maintain synchrony given that their biological clocks are noisy? I am currently involve in designing experiments (in collaboration with the LIMBS lab, Johns Hopkins) to collect and analyze data from humans performing simple sensorimotor tasks where they must cooperate to stay in sync. We plan to analyze the data with SysID. Please see a preprint of a relevant paper (click here).
Control Theory Where Agent(s) Rely on Noisy Clocks
No clock tells time perfectly. In the past, when timekeeping was not good enough for a particular engineering application, the engineers invariably built better clocks. They did not create algorithms to deal with noisy clocks. Thus, up to to now, it has been safe to assume perfection in a controller's timekeeping. Unfortunately, evolution has different constraints, and it is apparent an extension of control theory is needed to deal with the case that clocks are noisy. To date, surprisingly little has been done in this regard.
Pedagogy for Training Students on System Identification
I am interested in finding effective methods for training students in the art of system identification. Training students to use SysID is easier than might first be thought, because one can collect the experimental data needed from simulations. For training and testing purposes, using a simulated system to generate the data is desired, because the experimenter then knows the correct model ahead of time. Knowing the correct model allows an assessment of the suitability and effectiveness of the SysID methods. Moreover, an instructor can create a dataset from simulation, but withhold from students the structure and parameterization of the generating model, giving students the more realistic experience of not already knowing the answer. In my mind, the primary reason this feature is such an advantage appears comes from the fact that no large investment in a laboratory equipment and training is needed.
