Emotion Recognition And Robot Control Based On Piezoresistive Array Tactile Sensor

With the increasing demand for intelligent robots,human-computer interaction has gradually become a hot issue for scholars.In the future,the intelligence of robots must be developed in the direction of understanding user emotions.How to make robots have the same emotional cognitive ability as human beings,which requires constructing the robot's emotional recognition system,allowing the robot to correctly and naturally dig out the needs of human emotions without any instructions,and make corresponding feedback.Aiming at the limitation of emotion recognition methods such as visual,speech and physiological signals,this paper proposes a emotion recognition method based on piezoresistive array tactile sensor and applies it to family service robots.Combining emotion recognition with robot control,a kind of design is designed.Robot personalized service system.The working principle of piezoresistive tactile sensor is analyzed.It is proposed that the piezoresistive array tactile sensor is more suitable for human-computer interaction.The12x12 piezoresistive array tactile sensor is used as the data acquisition device of tactile emotion recognition experiment.Its structure and advantages Analyze and analyze the input and output relationship of the sensor,respectively,the relationship between the pressure and output voltage of the sensor,the relationship between the pressure and the sensor resistance,and the output characteristics of the sensor itself and hysteresis.analysis.Finally,the application background of piezoresistive array tactile sensors in the fields of Internet of Things,medical care,and robotics is introduced.The process of tactile emotion recognition has four steps: data acquisition,data preprocessing,feature extraction,and emotion recognition.In this paper,the piezoresistive array tactile sensor is used to collect the tactile emotion data,and the related data set is established.The data of the two tactile emotion data sets are preprocessed according to the characteristics of the noise,and the influence of the interference signal noise on the experimental results is removed.According to the characteristics of the tactile emotion data Emotional feature extraction is performed on two tactile emotion datasets respectively.The models of support vector machine,extreme learning machine and lifelong learning are briefly introduced,and a new algorithm model is proposed: lifelong learning algorithm based on extreme learning machine.The algorithm combines the extreme learning machine with the lifelong learning algorithm,and solves the problem that the lifelong learning algorithm does not take into account the nonlinear feature extraction,and has a good application effect on tactile emotion recognition.The emotion recognition effects of different classifiers are explored,and the extracted tactile emotion features are input into different classifier models for three sets of experiments.The first group of experiments used the extreme learning machine and the support vector machine to explore the effect of emotion recognition,but the results show that the emotion recognition effects of the two traditional machine learning algorithms are not very good;the second group of experiments is to separately identify the emotions of 14 gestures.The results show that the emotion recognition effects of different gestures are different,the emotion recognition effect of stroke gesture is the highest,and there is a certain relationship between gesture and emotion.Different gestures correspond to certain emotions,which will affect the experimental results;The third group of experiments used the life learning algorithm based on the extreme learning machine to explore the effect of tactile emotion recognition.The recognition effect is significantly higher than the two traditional machine learning algorithms.The experiment proves that the lifelong learning algorithm based on the extreme learning machine also has Good robustness.Finally,the author explores the emotional cognition and control of the robot,combines the emotion recognition result with the scene recognition and voice interaction function of the home robot,and designs a music preference selection strategy based on user emotion to control the robot to play the appropriate music.Improve the emotional state of the user and improve the cognitive and control effects of the robot.