Autonomous Cognition And Correction System For Robot Personalized Service Based On Emotional Information

Relying on the accelerated popularization of artificial intelligence technology,home service robots are rebuilding a scientific,comfortable and healthy lifestyle,giving people a happier and better life.In the process of robots gradually entering the home environment,in order to establish a harmonious way of human-robot interaction and improve the comfort and happiness of users in the home environment,people have put forward higher expectations for robots.In addition to the need for intelligence,it is hoped that robots can imitate humans’ ability to understand emotions.According to the user’s emotions,robot can simulate people’s emotional care thinking for task recognition and execution.The task planning of home service robots is effectively combined with emotion recognition technology to construct an autonomous cognitive method of robot service based on user emotion information,which is the focus of modern smart home service robot research.Based on the background of the smart space of the home environment,this paper designs an autonomous cognition and correction system for robot services with emotion as the core.Through the system,the robot can perform service task recognition and service content correction based on the user’s emotional state.In the process of service recognition,the corresponding personalized service items are selected according to the user’s personal preferences.The service content is tailor-made for users.After the end of each service,the robot learns the service knowledge and provides experience for subsequent service recognition.The main work of this paper is as follows:(1)A service robot emotion recognition system based on the user’s physiological signals is constructed.Firstly,the physiological signal is preprocessed.The filter is designed to remove the noise signal under the premise that the original signal information is not lost.Then the corresponding emotional features from the preprocessed physiological signals is extracted.Afterwards,two methods of ensemble learning and bi-directional long-short term memory recurrent neural network(Bi-LSTM)are used to construct the emotion recognition model.In the ensemble learning method,in view of the problem that the extracted features often contain redundant and weakly related features,the clustering and max-relevance and min-redundancy(mRMR)algorithm is used to perform preliminary feature selection and eliminate unnecessary features.After that,two machine learning algorithms,support vector machine(SVM)and K nearest neighbor(KNN),are combined for emotion recognition.In the Bi-LSTM method,the entire physiological signal is serialized.The emotional information contained in the signal is captured in the two dimensions of time and characteristics to realize the recognition of the emotional state.(2)Taking into account the importance of emotion in the process of robot service recognition,emotion recognition and case-based reasoning(CBR)are combined to establish a service cognitive model with emotion as the core in a smart space scenario.Firstly,the smart space scene data containing the user’s emotional information is encoded,and then the hash index and the nearest neighbor strategy are used for case retrieval,and the source case results are reused to realize the autonomous cognition of the service.After the end of the service,the service data is saved in the case library in an incremental manner to realize the enrichment and growth of the cognitive knowledge of robot services.(3)Facing the differences in interest preferences between users and the variability of user interest preferences,a robot personalized service recommendation system is established.On the basis of preliminary service recognition,further refine the service content.The collaborative filtering algorithm is used to filter content that meets the preferences of the target user based on the existing behavior of similar users.Then combined with the content-based recommendation algorithm to provide services on the premise of meeting the emotional needs of users,and make the final service item selection.Finally,the user’s emotional state is used as feedback to modify the robot’s service content.The robot’s understanding and understanding of interactive tasks is enhanced,reducing human participation and enhancing user experience.(4)Experiments are carried out for the multi-modal physiological signal emotion recognition method and the autonomous cognitive mechanism of robot service proposed in this paper.The experimental results show that the use of multi-modal physiological signals can achieve good emotion recognition results.Through the autonomous cognition method of robot service and the personalized service recommendation algorithm,the cognition and correction of the robot’s personalized service based on the user’s emotional information can be realized.And it can realize the self-learning of cognitive knowledge of robot service.In the smart space scenario of the home environment,the effectiveness of the robot service autonomous cognition and correction system with emotion as the core is verified.