【摘要】 在详细分析滑步式投掷铅球动作特性的基础上,设计了一种有效结合小波变换和模糊神经网络的运动员投掷力信息识别新方法。利用小波分解与重构的方法对信号进行了去噪处理,并采用小波系数的能量值作为运动员投掷力曲线的特征,将特征向量作为模糊神经网络的输入,对运动员投掷力曲线进行识别。经过比较实验验证,该算法既降低了噪声的影响,又在有效提取特征的同时减少了神经网络的运算量,提高了识别速度,具有较高的识别精度。更多还原

【Abstract】 A novel method for recognition of athlete’s throwing force is introduced in the paper, which is based on the motion analysis of gliding shot putting and combines the algorithms of wavelet transform and FNN. Using the wavelet decomposition and reconstruction method, the noise is restrained efficiently. In order to identify the throwing force curves of different motion phases, the signal features are extracted using wavelet transform method. The energy values of wavelet coefficients are chosen as signal features and then input into the FNN for recognition. The experiment shows that the method has high anti-noise ability. It not only extracts the features efficiently, but also decreases the burden of neural network. Therefore, the recognition speed is increased and recognition efficiency of neural network is improved.更多还原