A Quantitative Model To Evaluate Wrist-Rotation In Golf (End)
This method is able to provide details on the individual segmental contributions of the left arm to the ﬁnal swing movement. The calculated club head path is veriﬁed by comparing those obtained through video analysis. Inertial Measurement Units (IMU) were placed at the grip end of a golf club to measure acceleration and angular velocities with six degrees-offreedom for a golf putting training system. A putting robot capable of performing highly repeatable putts and independent measuring instruments as used to assess the accuracy of the sensor system. By using measurement theory, their system was able to provide the position of the club head to within mm and the orientation of the clubface to within 0.5◦. Golf training aids have also been proposed and which target speciﬁc problems faced by novice golfers. A swing guide is presented which aims to help the golfer coordinate the movements performed during a swing. The exercises the author recommends while using the device target the coordinated movement of the hips, shoulders, elbows, wrists, and the golf club. A mechanical golf swing training device is presented to help player sperfect their backswing and downswing movements. The training device helps players focus on the non-dominant arm and shoulder while keeping the swing in the proper swing plane. The proposed device helps players to develop muscle memory in their upper body to produce a smooth, consistent, and controllable swing. Another training device helps players to maintain proper right leg positioning during the backswing. Its purpose is to restrict the lateral movement of the right leg away from the target during the backswing while not hindering the forward movement of the legs during downswing and followthrough. The authors model the golf swing as a double pendulum system. Wireless inertial sensors are placed along the body and golf club to determine how closely the movements of the body follow predetermined motion rules. This is used as a quality measure for the golf swing. The authors deﬁne a physical model ofthe swing which accounts for thelength of the backswing, thewrist-cockangle, energy transfer during the swing, the swing plane, and club-head speed. Examples of other golf swing analysis systems can be found.
Though most of the above training systems are successful in introducing methods for golf swing analysis, the training aids can only be accessed at a specialized facility making wide spread deployment difﬁcult.Properties such as mobility and wearability make BSNs more promising for designing sport feedback systems. We take advantage of pattern recognition techniques in designing our training system to avoid the need for per-jointand complementary sensor deployment as required for kinematic analysis techniques.
Discussion and Future Work
Our quantitative model functions based on feature vectors from all sensor nodes across the network.With an exhaustive set of features obtained from each segment, this may yield in large volume of data for signal processingandcommunication.Theamountofdatarequired for signal processing, however, can be significantly reduced by the data reduction techniques described earlier. Since each sensor node partially contributes to the linear projections of PCA and LDA, it can combine local features using pre-obtained eigenvectors and transmit a single value for each trial to the base-station.
At this stage of our study, we process data ofﬂine. This is convenient for rapid prototyping and algorithm development. However, we have great suspicion that our algorithms for signal processing can be implemented and executed on the motes.
In this research, we focused on building our quantitative model for non-ideal movements due to wrist rotation. Evaluating this model for other types of incorrect swings requires controlling experiments for those types of errors. We plan to investigate this in future.
In this paper, we presented a system which uses body sensor networks for the purpose of golf swing training. We developed a quantitative model which can provide feedback on quality of movements for the purpose of training. The system architecture, signal processing methods and experimental results of the system were presented. The results demonstrate that our model is able to provide information on the quality of a golfs wing with respect to the angle of the wrist rotation.