2018 Fiscal Year Research-status Report
Development of the "Pilot's Individualized Learning using Objective Data (PILOD)" training method for increased training efficiency and objective assessment of manual flight skills
| Project/Area Number |
16K21002
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
エントジンガー ヨルグ 東京大学, 大学院工学系研究科(工学部), 助教 (60600327)
|
|---|
| Project Period (FY) |
2016-04-01 – 2020-03-31
|
| Keywords | Pilot training / Personalized training / Pupil size / Workload evaluation / Meta-analysis / Objective feedback |
|---|
| Outline of Annual Research Achievements |
As planned, I have finished a meta-analysis of data gathered from all subjects who took part in the experiments so far. I expanded the analysis from mostly performance metrics to include automated evaluation of physiological measures as well. These results showed a clear overall improvement of the objectively measured, and now also automatically analyzed skills with training. I presented these results at international conferences.
I developed new training tools as well as added features to the flight simulation to provide specific exercises depending on each trainee's individual analysis results. This will enable me to actively close the PILOD feedback loop by better customizing the training, instead of mainly relying on oral or visual feedback.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The analysis of data is more and more automated. The system requires less expertise and user interaction than before, and can therefore handle more experiment data with less effort. The increased number of additional training exercises available also is expected to improve the effectiveness of PILOD. We also carried out several experiments with subjects holding a piloting license as a comparison to the general population.
At several conferences I attended, other researchers mentioned that they are thinking about a similar approach, but it seems most of them are still in an orientation phase. These is much interest in the actual results I present.
|
| Strategy for Future Research Activity |
I am discussing with a third party about possible collaboration on experiments or sharing data for analysis. This could verify whether the data recording and analysis developed in the PILOD project are suitable in actual environments. Since the (trainee) pilot population is expected to be more homogeneous than the general population that took part in our experiments so far, some discriminating measures we found might not prove useful, and we will have to tune or change the analysis to find meaningful results.
|
| Causes of Carryover |
I recently hired research assistants who will help me with meta-analysis and the application of machine learning to find new possible metrics to evaluate safety and risk from the flight and training data. We also expect a large amount is needed for the repair and maintenance of our flight simulator, which had been postponed as we were using it for many experiments. Finally, publishing the results in a journal or on conferences will require a significant part of the budget.
|
