Transfer of training in general V.R. environments

1
Transfer of training in general V.R. environments
PPS Project
Benjamin Bollmann, Lukas Friedrich
ETH Zürich
2
Training in Virtual Reality (1)

• VEs as a training tool Cheap, secure,
  repeatable, realizable, etc.
• Further advantages free setting of the
  environment parameters, control over the
  development of the training, easy monitoring of
  the trainee response.

3
Training in Virtual Reality (2)

• VE as a game format increased motivation (e.g.
  young population) better learning
• Why should we still train in real systems?
• 1) Technology limits real environments hard to
  reproduce (computing power, mechanical
  feedbacks, etc).
• 2) Transfer of training Does transfer of
  training always exists from a VE to the real
  world? Always effective to train in VEs?

4
Transfer of training

• The term transfer of training concerns the way
  in which previous learning affects new learning
  or performance
• Positive transfer from task A to task B learning
  task A improves performance in task B. Negative
  transfer learning task A interfere with the
  procurement of task B.
• Factors affecting the transfer Similarity of A
  and B, practice, motivation, self-management,
  etc.

5
Measuring the transfer of training

• Ideally An experimental group trains the task on
  the VR-training device. A control group trains in
  the real world. Comparison of the result through
  a test in the real world. Results of the
  experimental group Results of the control group
  effective VR-training device
• Practice we talk about a positive transfer of
  training if the results of an experimental group
  are better in a post-training test than in a
  pre-training test.

6
Experiment 1 steadiness tester (1)

• Task guide the loop along the wire as quickly as
  possible without touching it (non preferred
  hand).
• 3D virtual environment via head-mounted display.
• 3 groups (total 150 people). Training
• 1) group RW (control) 8x in the real world
• 2) group VR (experimental) 8x in the VE
• 3) group NP no training

7
Experiment 1 steadiness tester (2)

• Results of the post training test
• More errors in the NP group. No significant
  difference between RW and VR group.

8
Experiment 1 with interferences (1)

• New test equal repartition of the control and
  experimental group in 2 new groups a motor and a
  cognitive group.
• Motor group task same as before but this time as
  typing a key 2 times per second with the free
  hand.
• Cognitive group task same as before but this
  time as listening to pre-recorded audio sounds
  (40 fruit words).

9
Experiment 1 with interferences (2)

• Results of the test
• More errors in the motor group. The VR group
  shows better results than the RW group for both
  motor and cognitive task.

10
Experiment 2 pouring movement (1)

• Task 4 acquired brain injury patients are
  trained for a particular motion corresponding to
  a pouring movement with a cup.
• The VE consists of a 3D program on a desktop
  computer. Patient must follow a teachers
  trajectory on the screen with a 3D mouse.
• No control group. Training
• S1 32x
• S2 16x
• S3 16x 16x (mid-way and final test)
• S4 16x 16x (mid-way and final test)

11
Experiment 2 pouring movement (2)

• 4 different tests virtual without teacher, real
  world pouring, Fugl-Meyer Test of Motor Recovery,
  Emory Test of Upper extremity.
• Results of the tests

12
Experiment 3 Spatial information (1)

• Task learn in a VE how to escape from an unknown
  building via the fire exit
• 10 physically disabled children, mean age 9
  years
• VE-system program in the format of a game
• Training method navigate through the virtual
  environment of the building and find the
  fire-exit signs

13
Experiment 3 Spatial information (2)

• Results
• Significant results
• All children find the way out
• Clear positive transfer of training
• Especially children benefit from that type of
  training

14
Experiment 4 LEGO biplane model (1)

• Task build up an LEGO biplane model in reality
  after training on a VE-system
• 15 students (20 years) split up in 3 groups
  concerning their abilities
• 3 forms of training
• VBB System with force feedback
• VBB System without force feedback
• no virtual training

15
Experiment 4 LEGO biplane model (2)

• Results
• Force feedback is important for training in VEs !

16
Discussion (1)

• Training effect dependent on
• complexity of the experiment
• degree of realistic representation
• Interesting difference under a interference (exp.
  1)
• movement is learned more automatically
• VE training is more difficult without
  haptics better training, better results

17
Discussion (2)

• Improvements not only in particular motor skills
  but also in general motor skills (exp.2)
• careful experiment shows differences between
  the subjects dependent on personal parameters
• Spatial information through VE-training program
• people can orient themselves in VR
• especially benefit for children, gaming factor of
  VE-system
• benefit for disabled children/people for safety
  purposes

18
Discussion (3)

• Higher learning performance with force feedback
  (exp.4)
• more realistic
• association to objects increases
• perceptual better prepared
• In general three components of training
• cognitive training
• perceptual training
• motor training
• Force feedback addresses all three components

19
Conclusion

• 4 different experiments Positive transfer
  of training.
• Unexpected results hard to generalize
• Potential with young generation
• Encouraged to put it more into practice

20
Thanks for your attention!

• Questions?