基于传感器技术的儿童益智游戏陪玩机器人设计

The design of children’s educational game accompanying robot based on sensor technology

随着科技的进步,儿童陪伴机器人的种类、功能也越来越多。但目前市场上的儿童陪伴机器人存在互动性差的特点。为了解决这些问题,更好地将儿童陪伴机器人融入到学前教育中,更大程度地激发学前儿童的思维、动手能力,研究基于传感器技术设计了具有新算法的儿童益智游戏陪伴机器人。该机器人首先利用视觉和听觉系统获取数据信息,然后利用模型算法对模糊神经网络(Fuzzy Neural Network,FNN)进行优化。然后利用递归神经网络(Recursive Neural Network,RNN)、卷积神经网络(Convolutional Neural Networks,CNN)与模型算法进行对。最后结果显示,在不同语言和环境背景下,FNN的语音准确率平均值、F1值和动作捕捉准确率平均值和F1值分别为78.51%、71.86%和61.83%、56.24%。均为三种算法中的最优值。实验结果表明,研究设计的儿童益智游戏陪伴机器人具有较高的语音识别能力和动作捕获能力。为儿童机器人的发展提供了一种新的思路。

With the advancement of technology,there are more and more types and functions of children’s companion robots.However,the children’s companion robots currently on the market suffer from poor interactivity.In order to solve these problems and better integrate children’s companion robots into preschool education to stimulate preschool children's thinking and hands-on abilities to a greater extent,the study designs a children’s educational game companion robot with new algorithms based on sensor technology.The robot first uses the visual and auditory systems to obtain data information,and then uses the modeling algorithm to optimize the Fuzzy Neural Network(FNN).Then Recursive Neural Network(RNN)and Convolutional Neural Networks(CNN)are used to pair with the model algorithm.The final results show that the mean value of speech accuracy,F1 value and mean value of motion capture accuracy and F1 value of FNN are 78.51%,71.86%and 61.83%,56.24%respectively for different language and environment context.All of them are the best values among the three algorithms.The experimental results show that the research-designed children’s educational game companion robot has high speech recognition and motion capture ability.It provides a new way of thinking for the development of children’s robots.