Application Research on Sport Skill Learning Using Quantum-Inspired Differential Evaluation with back Propagation Neural Network

In this era, athletics is becoming a major developing area through the large number of training with Artificial Intelligence (AI) and Internet of Things (IoT). The training universities train the sport talents and encourage them to endeavor for glory of the country. However, in existing several AI devices has been utilized for sport skill learning including Gold SA (Genetic Algorithm with Simulated Annealing) optimization but struggled with low convergence and poor performance. In this paper, Quantum-inspired Differential Evaluation with Back Propagation Neural Network (QDE-BPNN) is proposed to increase performance with high convergence. Initially, by using sensors such as fitness trackers are utilized to collect the real time data from the athletes. Then, sensor data is transmitted and further preprocessed with min-max normalization. After that, the preprocessed data is fed into QDE optimization to evaluate weight and threshold values for the network. Finally, the output from optimization further processed with BPN network to evaluate how much sport skill is learned by athlete. From the results, the proposed QDE-BPNN established better results than existing IoT assisted with Physical Activity Monitoring Device (IoT-PAMD) including accuracy, precision and recall, Fl-score of 98.5%, 96.7% and 95.8%, 93.9% respectively.