Here’s an interesting project that was recently presented at the Media Computing Group at RWTH Aachen University. The wearable snowboarding assistant (pdf, 2.7 MB) makes use of a number of vibrotactile actuators worn at various points on the body to give real-time feedback about whether the user is performing within optimal ranges of motion and posture. The training system, which is still under development, will apparently use sensor recordings from expert snowboarders to create a reference signal of optimal weight distribution, joint angles, and other parameters. The trainee’s vibrators will be activated when the sensors on his or her body detect a deviation from the optimal input.
This is an example of feedback that may be called “augmented”,Â where a channel of information is overlaid on the normal sensory inputs of a task. Techniques such as these have been shown (PS, 24 MB) to be useful for acquiring sensorimotor skills. A feedback signal can be generated by subtracting the input from sensors monitoring the task from a sensor recording made by an expert while doing the same task. Vibration is a very effective way of augmenting feedback, because it is inexpensive, somatic, and has a low attentional requirement. I’m very excited to see how the wearable snowboarding assistant develops.
For some background on augmented feedback, see this excellent book.