One of the main problems sports players have to deal with is the difficulty of performing some motions with precision and reproducibility. I am thinking about tennis serves, free-kick shots in soccer or long-distance shots in basketball. These motions involve throwing a ball with a very measured amount of strenght and with a very precise direction in order to achieve the optimal result. The only way to achieve the necessary accuracy is to practice these movements to exhaustion so that the body can get used to them and perform them in an almost instinctive way. That is what we call muscle memory, the ability of our neuro-muscular system to repeat a motion without us really thinking about how exactly we are performing it.
But muscle memory is not easy to establish. The amount of repetition needed to achieve almost-perfect muscle memory is extremely high. As an example, the best shooters in the NBA shoot hundreds if not thousands of shots a day, every day, for years, before they become that good.
One of the problems with establishing the muscle memory is the lack of feedback when we do something right or wrong. The athlete can focus too much on the effect of the motion (whether the ball goes in or not), or the "feeling" of it, instead of the technique, often leading to developing bad habits or being inconsistent.
Haptic feedback, if well tuned and coupled to a system to measure optimal motion, could be used to give sensory feedback when the movement is being performed wrong. It could even be adapted so that the feedback told the athlete exactly what they are doing wrong, by giving the sensory imput into the part of the body that is not being used properly. For example if a basketball shooter tends to open their shooting elbow too much, the feedback would be on the elbow. Then if they start closing it too much they would receive the feedback on the inner arm.
The haptic feedback could also be used for the correct motion, making the brain associate the correct motion to a positive reinforcement may increase muscle memory consolidation and lead to faster improvement.
I find this a fascinating topic and I look forward to discussing potential implementations on this. For me the crucial thing would be how to measure the accuracy of the movement to be able to give appropriate feedback. Can we get some ideas on that?