Research On Humanized Behavior Decision Of Service Robot

With the continuous development of science,technology and economy,the aging of the population,and the protection of the daily life of the elderly and the disabled,service robots have played an important role.The demand for service robots is gradually increasing,and the service level of server people is expected to be higher.When serving the elderly,the service robot not only completes the task instructions of the elderly,but more importantly,it embodies the intelligent analysis and decision-making of robots similar to human beings.Because of the personality characteristics of the elderly(such as inconvenient physical movement and poor language expression ability),the information obtained from the task instructions is often vague and uncertain.The service robot needs to use these fuzzy information to simulate the logical thinking of human beings for reasoning and decision-making,so as to infer the real intention of the elderly.In order to solve the problem that the task instructions received by service robots are vague and not specific when serving the elderly,this paper constructs a hierarchical framework structure of humanoid behavior decision model,which makes a vague and not specific task clear and specific by simulating the way of human thinking reasoning and decision-making.The paper includes the following parts:(1)By simulating the process of human cognitive behavior decision-making,a hierarchical framework of human behavior decision-making model is established.According to the characteristics of perception layer,evaluation layer and decision-making layer in the hierarchical framework,the model structure of each layer is established,and the fuzziness and uncertainty of human cognitive behavior are simulated by the modeling methods of production system,Bayesian network and human-computer interaction respectively.And the flow chart of reasoning algorithm of human-like behavior decision model is given.(2)In the perception layer,when reasoning and judging the task objects in the received task instruction information,a reasoning method similar to human brain cognition is adopted in the production system structure,and the perceived state variables are input to the evaluation layer.(3)In the evaluation layer,according to the current state information and events output by the perception layer,and referring to the experience and prior knowledge contained in the knowledge base and the structure and parameters of Bayesian network,the current situation information is understood and judged,and the possibility of various events and situation conditions appearing in the next step is evaluated.(4)In the decision-making layer,through the situation information evaluated by the evaluation layer,the varieties and attributes of specific task objects can be obtained according to the rules of probability.Then,according to the prior knowledge in the knowledge base,more information can be obtained with the least number of interactions in the way of human-computer interaction inquiry to complete the final decision,and the final results are feed back to the perception layer and the evaluation layer to update the information stored in the comprehensive database and the structure and parameters of Bayesian networks in the knowledge base.(5)Finally,through the simulation and analysis of different situations and information of different types of tasks under the same type of tasks,the feasibility of the reasoning algorithm of this kind of humanoid behavior decision model is verified.