摘要
为了使机器人在嘈杂环境下快速识别音乐并跟随节拍做出恰当的舞蹈,降低人工分析与调试动作的工作量,提出一种基于声纹特征峰值点进行情感分析,并根据峰值点信息自动生成机器人舞蹈动作序列的系统,实际部署并实现了一套速度快、准确度高的音频特征提取与搜索方法。机器人通过对音频能量值与连续性等信息进行处理,排列组合音乐特征参数,形成动作数据文件,确定姿态与运动轨迹,完成机器人自动生成舞蹈动作。仿真实验结果表明该方法显著提高了嘈杂环境下机器人识别音乐的速度和准确性,机器人能够根据峰值参数演绎出稳定、流畅、情感恰当的舞蹈,具有较强的可行性。
In order to enable the robot to quickly recognize the music and dance following the beat appropriately in noisy environment,so as to reduce the workload of manual analysis and choreography,a system based on peak points of voiceprint features for sentiment analysis and that can automatically generate robot dance action sequences based on the peak point information is proposed.In addition,an audio feature extraction and search method with high speed and accuracy is actually deployed and implemented.The robot performs processing of audio energy values and continuity and other information,arranges and combines music feature parameters,forms action data files,determines posture and motion trajectories,so as to complete the automatic generation of dance movements by the robot.The simulation experimental results show that the proposed method can significantly improve the speed and accuracy of the robot in recognizing music in noisy environment,and the robot can perform a stable,smooth and emotionally appropriate dance according to the peak parameters.Therefore,the proposed method is highly feasible.
作者
赵潇帆
彭熙
常亚楠
郑世珏
ZHAO Xiaofan;PENG Xi;CHANG Yanan;ZHENG Shijue(School of Computer Science,Central China Normal University,Wuhan 430079,China)
出处
《现代电子技术》
2023年第15期84-88,共5页
Modern Electronics Technique
基金
国家自然科学基金青年基金项目(61702210)。
关键词
舞蹈机器人
声纹情感分析
韵律识别
舞蹈生成系统
自动化
机器人架构
音频峰值转化
自动编排
dance robot
voiceprint emotion analysis
rhyme recognition
dance generation system
automation
robot architecture
audio peak transformation
automatic choreography