Service Robot's Structure And Achievement Of Smart Space

In our country we did not research robot until recent decades.However, robot technology has had a soaring development,and some new achievements and products have been issued since recent years.This paper research some kinds of smart space technology in the world and introduces the technology into indoor service robot field because of relative disadvantanges of robot research in China.This paper is writed to resolve a problem that nowadays service robot can not achieve service tasks accurately in unstructured and time-varying indoor environment such as classrooms and families.We focus on the following key technologies and finally present new outlooks.This paper discusses detailedly how to design the smart space.First,in order to make readers better to understand this paper, the conceptions and characters of pervasive/ubiquitous web and smart space technology are introduced.Then the author presents the global structure of smart space:in order to supply global entironment information,the author structures the global grid map by indoor entironment's characters and using image processing technology; In order to achieve real-time Information exchange, the author structures information communication by Bluetooth technology;Finally, in order to help robot position and supply entironment information, the author designs a new landmark by artificial landmark's theory. After the demonstration, the global grid map can describe indoor entironment detailedly and totally;the designed artificial landmark can not only fulfil sufficiently its theory,but also be easily detected and recognized,which could supply sufficient entironment information and help robot position.This paper discusses detailedly how to achieve the smart space.First,the author introduce service robot's sensor parameters.Then simultaneous localization and mapping is focused on:In order to achieve state-updating and postion-forecasting, the author structures model of movement,observation and tatus updates by extend Kalman filter theory; In order to achieve obstacle avoidance and navigation, the author structures local grid map by ultrasonic theory and ultrasonic and infrared sensor technology; In order to resolve data association,which is the problem of determining whether or not two sensor measurements observed at different places and time, two features in different partial maps, the author uses multiple hypothesis tracking-MHT theory.Finally,in order to fuses information effectively and reliably, the author proposes a new multi-sensor information fusion technology by the D-S evidence theory and BP neural network theory. After the demonstration,the structured local grid map obtains a lot of detailed entironment information; such technology fuses multi-sensor information real-time, effectively and reliably.This paper discusses detailedly path planning and obstacle avoidance.First, the author resolve robot's path planning. the author structures entironment's map by grid in order to achieve path planning and obstacle avoidance; Path planning is achieved by reverse D*algorithm and grid map;Two-level obstacle avoidance strategy is using to achieve robot's real-time dynamic obstacle avoidance by ultrasonic sensors,which can make robot achieve tasks without collision.