Research on Flexible Reasoning of Service Tasks for Service Robots Based on Intelligent Space

It is an important criterion to measure the intelligence of robot that how to use the existing service information in intelligent space to better perceive the intentions of users and provide the required services for users in dynamic environment. In order to improve the service quality of service robots, service reasoning method based on machine learning algorithm and user intention recognition based on Multi-Entity Bayesian Network(MEBN) in dynamic environment are presented. These methods can deal with the drawback that service robots often ignore the context information generated by users and the rules of service reasoning are rigid when providing services to users in home environment. Firstly, the context information model is established for the existing user information in the intelligent space, and then the services required by users in a specific scenario are reasoned based on the K-Nearest Neighbor(kNN) algorithm. User's intentions may be complicated or varied because they will be affected by the dynamic environment, finally the MENB technology is used to incorporate logic theory into Bayesian reasoning for flexible rule reasoning so that the user's intention recognition is realized to provide satisfactory services to user. The simulation results show that the proposed method can reasoning the services required by users in dynamic environment, effectively improve the service reasoning ability of the home service robot, and make the human-robot integration environment more harmonious.