Research On Navigation Strategy Optimization For Robot Intelligent Service

With the growth of social demand and the development of artificial intelligence technology,service robots have entered life to provide users with convenient services.The intelligent navigation method of the robot is the premise of completing the service task,and how to improve the navigation intelligence has become an urgent problem to be solved.Aiming at the problems of task planning efficiency,object search accuracy and comfortable service in the process of robot task execution,this paper takes intelligent service as the guide to explore intelligent navigation methods of robots.The contributions of this paper are as follows.(1)A task sequence generation method based on probabilistic reasoning is proposed.The chaotic indoor environment increases the difficulty for robots to obtain information and reduces the efficiency of task execution.Indoor environment knowledge covers the semantic information of rooms,static items and dynamic items,as well as the spatial position relationship of items,which has a guiding role for task sequence generation.In order to make the robot complete the planning task efficiently,two probability update methods are proposed to calculate and weight the probability information of item position respectively.The probability update function is designed so that the robot can update the co-occurrence probability after finding the target item and guide the subsequent search.A probabilistic update method is used to guide the JSHOP2 planner to generate plausible task sequences and decompose them into action sequences for robot navigation.(2)A hybrid search strategy to guide robot search is proposed.The hybrid search strategy includes a search strategy based on prior knowledge and a probability update model as well as an active visual search strategy.The distance function is designed to calculate the distance between the robot’s current position and the target position,and combined with the probability update model to select the optimal navigation path.Some static objects in the room will be moved by the user and change their position.To solve the problem of the uncertainty of the position of static objects,an active visual search strategy based on conditional random fields is improved to allow the robot to autonomously explore the room and find the target faster.(3)A comfortable service mechanism for service robots is proposed.The comfortable human-computer interaction ability and navigation ability of robots are very important.Robots need to identify user behavior and state to determine the comfortable interaction distance.The comfortable interaction distance is determined by taking the joint velocity information and position information as the input of the Mamdani fuzzy inference system.Considering the comfort of users and the efficiency of task execution,the speed of humanrobot interaction and searching for objects should be different for service robots.For this purpose,we set the micro-speed adjustment method(comfortable approach speed)and the macro-speed adjustment method(robot running speed).By using user behavior,user expression and joint position information as inputs to Mamdani fuzzy inference system for determining comfortable approach speed.At the same time,the human-computer interaction interface is set so that the robot can perform macro-speed control in all indoor scenes.The comfortable interaction distance adjustment strategy and the comfortable speed adjustment strategy are combined to realize comfortable navigation in task execution process.Guided by intelligent services,this paper optimizes the intelligent navigation strategy by combining environmental cognition ability and reasoning ability of service robot.The methods in this paper improve the ability of task planning,object search and comfort service,have important theoretical and practical significance for the optimization of robot navigation strategy.