A Service Robot Instruction Parsing Method for Action Sequence Generation in Intelligent Space

Service robots face challenges in difficult task analysis, poor environmental adaptability, and unfriendly human-machine interaction when performing household tasks. In order to accurately understand user intent and plan service steps suitable for the current environment, an intelligent parsing methods for service robot instructions in intelligent spaces is studied, to enhance the cognitive abilities of robots for service tasks in a home environment. Firstly, starting from the keywords in instructions, an intent recognition model based on a gating mechanism is proposed to enhance the robot's cognitive ability towards user instructions. Secondly, a sequence text generation mechanism is proposed, using service strategies as intermediate states, to assist to generate action sequences for the robot. Additionally, a strategy correction mechanism based on intelligent space ontology interaction is employed by integrating multimodal perception between the robot and the environment, to generate service strategies that are most suitable for the current environment. Finally, a task planning module based on domain and problem descriptions is integrated to generate executable action sequences for the robot, thereby improving the quality of service task execution. Experimental results demonstrate that the proposed method maintains a friendly interaction while accurately understanding complex user instructions, and ultimately generates reliable action sequences that the robot can execute.