Speed-Distance-Time Experiment: Implement Operation-, Action- and Task-level feedback

[bastianilso]

For the upcoming experiment, we are trying to understand what microinteractions create effective training environments in exergames. For that we need a range of different behaviors we can test against one another.

Here is a summary of the proposal scope:

Involved Entities:

• Cursor
• Mole
• Wall

Levels:

• Operation-level (Cursor move, issue #240 )
• Action-level (Mole Hit, issue #256)
• Task-level (Progress Wall, see issue #241)

Information Channels:

• Speed
• Travel Distance
• Time

One obvious limitation is that we won't be able to test every combination of entity × level × information channel. However, the plan is to test each information channel and then choose specific pairs of entities and levels. The argument being that although all combinations are possible, not all combinations necessarily make sense. What makes sense, has to do with how the actors relate to the user's interactions at any given level. For example, using the mole for operation-feedback using speed as information channel, while technically possible, makes less sense because the player is manipulating the cursor in real-time, not the mole.

Thus the following mappings are used:

• The player moves the cursor in real-time, therefore the cursor movement is targeted for operation-level feedback (issue #240)
• The player hits the mole, in order to complete the hit action successfully, therefore the mole is targeted for action-level feedback (see issue #256)
• The wall contains all moles the player is supposed to hit during the target-to-target pointing task, therefore the wall is targeted for task-level feedback (see issue #241)

Status:

• @mrvilix06 has implemented the base performance manager and is finalizing the action-level feedback.
• @tcaruchet will be looking into the wip/TrajectoryAssistantAndFeedback branch, to recover and refactor the operation-level feedback made there. @tcaruchet will coordinate the implementation with @mrvilix06's performance manager.

Remaining Todos:

• Implement the other information channels travel distance and time (making it possible to switch between them).
• Implement a progress bar UI on the wall and expose it as a target for performance manager's feedback.
• Implement support to switch between action, operation, task and combined feedback from treatment patterns via the PERFORMANCEFEEDBACK parameter.

[bastianilso]

note to self:

Speed at the "Operation"

• Principle: We may only make use of the information available within the operation. In this case, we can use the human-motor response rate to create a moving range of information available to operation feedback.

Speed and moving averages:

• Moving averages seemed like a nice idea to create a middle ground from which we could define "worse" and "better" feedback. If users improved, the moving average would improve along with them, providing more challenge. However, at the operation-scale, using a moving average created by the action become problematic because of the wide range in which speed varies, making speed unpredictable. Also, it breaks the principle stated above.
• Alternative: Use a "max-observed-speed" principle. Define an average of the 10 maximum observed speeds. Define lowest positive feedback as half that speed. We can define max-observed-speed and half speed better at both operation and action levels separately.

other approaches might result in normal distribution curves and upper/lower quartiles, which is a can of worms im not sure i want to open, if a simple option can do the job.

[MadsKastberg]

Here´s 4 sound effects that I made that you can use as you like. Feel free to dump them if they dont match your desired goal! These are 4 very different sounds, but the common theme is that they should all loop seamlessly.

Seamless audio loops.zip