Research On Social Behavior Model Of School Reception Service Robot

As a kind of service robot with social intelligence,the ultimate goal of social robot is to be able to interact with human naturally and effectively in human environment.The school reception scene is a typical social interaction scene,which will generate a series of complex social interaction behaviors.At present,it is difficult for service robots to deal with the interaction intention of multiple interactive objects at the same time in the school interaction scene.The task-driven behavior generation method makes its behavior solidified.The single-line processing and one-by-one task processing mode is inefficient in complex scenes.Therefore,this thesis takes the school reception scene as the object to carry out the relevant research on the social behavior of service robots:⑴The key problems of social behavior of service robots are analyzed.Based on the social interaction scene of school reception,the relevant technology is discussed,and some problems faced by the technology are analyzed,and the common technology methods of robot social behavior are summarized.⑵The social behavior model of robot is constructed.This thesis proposes a social behavior model for service robots.The main work includes:(1)Based on hierarchical task network,social intention is decomposed into four levels: behavior,time series,action,and actuator,which are transformed into corresponding social behaviors.(2)According to the six-dimensional emotion vector model proposed by Ekman et al.,the robot emotion module is constructed,and two kinds of emotional stimuli,namely satisfaction state and user evaluation,and emotion attenuation mechanism are introduced.(3)Construct the robot demand structure according to Maslow’s demand behavior theory,and design the robot’s internal demand generation mode.(4)Based on emotions,roles and needs,the thesis constructs the weight of social behaviors of service robots to help robots judge the importance of behaviors.(5)Through the simulation test and analysis of the emotional module and the demand module in the robot social behavior model,the thesis verifies the effectiveness of behavioral weight construction.⑶A scheduling method of robot social behavior is established.The key contents include:(1)Aiming at the situation that the current task of the existing service robot is easily interrupted by unimportant user intentions and cannot be recovered after being interrupted,a robot behavior model with interrupt mechanism is designed.(2)A behavior concurrency mechanism based on the free flow model is established,which combines with the robot social behavior model to ensure that when there are multiple candidate behaviors,the optimal behaviors are executed simultaneously with the non-conflicting sub-optimal behaviors.(3)The Little Fat robot platform in the laboratory is used to simulate and verify the above research in the school scene,which proves the feasibility of the method.⑷The behavioral system of social service robot in the school reception scenario is developed.(1)Establish the system structure by using software engineering technology.(2)Construct the social behavior model of service robots based on ROS,and establish the social behavior scheduling method.(3)The above research results are integrated into the entity robot "FABO",and the visualization technology is applied to complete the entity experiment,which verifies the reliability,availability and robustness of the system.