Research And System Implementation Of Audio-visual Interactive Technology For Domestic Robot

By the end of 2017,the population over the age of 60 in China amounted to 240 million 900 thousand,accounting for 17.3% of the total population in China.The population over the age of 65 was 158 million 310 thousand,accounting for 11.4% of the total population in China.The aging society has arrived.With the development of artificial intellige technology,more and more companies and organizations are engaged in development of robots,and robots of various forms and functions have appeared one after another.In the future,robots will gradually undertake some repetitive and heavy work,and become the right-hand in the daily life of the elderly.It is found that voice interaction and target detection technology are used more frequently in human-machine interaction.Voice,as a convenient,fast and efficient way,will undoubtedly become an indispensable means for people to communicate with the robot,and the target detection is equivalent to gave a pair of "eyes" to the robot to make it more intelligent.Therefore,it is necessary to research voice interaction and target detection technology and design and implement a human-machine interaction system for domestic robots based on these two technologies.The existing human-machine interaction system for domestic robots is not rich in functions.Some systems can only recognize specific voice control instructions and can not identify the instructions with similar meaning.Some systems can only respond by matching people's questions and local problem templates,they have single function and poor expansibility.Although some systems can carry out target detection,the target detection effect under dim light,occlusion and truncation is very poor.In order to solve these problems,this paper researches the speech interaction and target detection technology based on Turtlebot robot,and a human-machine interaction system for domestic robots is designed and implemented.This paper designs and implements the human-machine interaction system for domestic robots.First of all,in order to better realize the robot voice control function,the system joins the corresponding relationship between the input instruction and the control instruction library,and the user can control the robot according to its own habit.Moreover,this system adds the voice control commands to the userwords and uploads them to the server to improve the recognition accuracy.In order to realize the function of human-machine dialogue,this paper creates the common problem library and the real-time problem library respectively,and defines their contents to meet the needs of the daily chat of the home old people.Secondly,the control strategy of the robot target tracking motion is expounded.The target is detected by the computer gpu,and the target detection module can better serve the home old people by training dark light and occlusion data set and the removal of other objects outside the pedestrians.Finally,Based on the related technology,the human-machine interaction system for domestic robots is implemented.In order to verify the performance of human-machine speech interaction and target detection proposed in this paper,they are tested many times on pc and the robot platform.The result proves that the robot nursing system proposed in this paper is highly reliable and practical.