Pharmacy Intelligent Service Robot Based On ROS

With the development of society and the introduction of relevant national policies,more and more industries began to transform towards intelligence,and various artificial intelligence applications appeared.With the joint efforts of researchers and developers,smart medicine has gradually appeared in all kinds of medical systems.Under the influence of the current Internet,advanced equipment and technology make the current medical system develop towards a more intelligent and information-based direction.The development of service robots is one of the reasons to promote human beings to enter a more intelligent life.At this stage,service robots have a wide variety of products and a wide range of fields,which can assist human beings in their work.With the decrease of medical professional service talents and the increase of the number of offline pharmacies,the demand for professional sales pharmacists in pharmacies is gradually increasing.At present,in the robot development platform,robot operation system(ROS)has attracted the attention of today’s robot developers with its excellent structure system and community ecology.As a technology integrating multi-disciplinary knowledge,speech recognition technology will enable robots to recognize human language.The more accurate the speech recognition is,the more the robot understands people’s intentions,so that it can give a correct response.In this trend,more and more robots and even intelligent devices have realized the function of speech recognition.Based on the above background,this paper designs and implements a service robot that can provide help to the masses in the pharmacy environment,including the functions of receiving users’ consultation and shopping guide.This paper focuses on speech recognition algorithm,analyzes and discusses several network models suitable for this research,and designs and improves speech recognition algorithm according to the framework of speech recognition algorithm at this stage.The main innovations of this paper include:(1)The method of unifying audio data is introduced in the process of speech recognition.Before recognizing the audio files in the data set,the data is cleaned by limiting the speech length and discarding the illegal characters in the annotation file,and then the error caused by different people’s speaking habits is reduced by unifying the audio frequency,so as to improve the robustness of the system.(2)In the process of acoustic model training,the network architecture is improved to ensure that the size of the model is compressed as much as possible on the basis of high recognition accuracy,so as to compress the operation time of the model to the greatest extent and improve the human-computer interaction experience.In the construction of network framework,RNN network architecture more suitable for sequence tasks is selected as the backbone of the whole speech recognition algorithm,and GRU is selected to form the algorithm structure after comparing various RNN models.Finally,the CTC loss function is used to optimize and adjust the whole model,which not only realizes accurate identification,but also reduces the time overhead.(3)Based on the research of the above algorithms,this paper uses the ROS robot development platform to design the complete function and framework of the pharmacy intelligent service robot system based on ROS,build the database,complete the functional modules of the system,and finally realize the pharmacy intelligent service robot system based on ROS.