神经网络在仿生机器马速度预测中的应用

Application of artificial neural network in the virtual speed prediction of mimetic robotic horse

本文首次提出将神经网络应用于生理信息的融合,针对仿生机器马运动过程中,系统如何根据使用者自身的身体素质,实时监控运动过程中人体生理信息变化,自动变换运动方式,从而获得最优锻炼效果的问题,本文提出了用BP神经网络来实现最优虚拟速度值的预测,通过改变虚拟速度使骑乘者改变骑乘姿势,从而获得最佳的运动量,经过实验表明,控制效果可以满足所要求的精度,训练后的模型可以用于速度预测。

The article unprecedentedly asserts the application of neural network in data process of bio-information. To deal with the issue that in the movement of bio-mimetic robotic horse, how to monitor real time bio-information in the process and automatically adjust riding gesture based on the rider’s physical conditions to obtain optimal results, BP neural network is proposed to predict optimal virtual speed, which is changed to adjust rider’s posture to obtain the optimal sport quantity and quality. Experiment demonstrates that the result meets required precision. After-excise model can be applied to speed prediction.