The Design And Implementation Of Interactive Autonomous Navigation Robot Platform Based On ROS

In the 21st century, the applications of automation technology and information technology in robotics are becoming mature gradually, which prompts the robot to enter the era of intellectualization. Intelligent robots are coming to our lives. Researches of intelligent robot are valuable, but there are some exiting problems:first, there are too many forms and functions of robots, making the selection of robot hardware and software difficult to meet specific research demand; second, in terms of natural human-robot interaction, most systems are developed with single sensing technology, which makes the applications limited; third, in terms of autonomous navigation, a lot of new deliverables are being applied, but most of them lack of natural human-robot interaction abilities. This paper researches the above problems from interaction perspective, and gets three major results as follows:Firstly, regarding the problem that it’s hard to satisfy specific needs of intelligent research, this paper proposes LAMDA-ROB-2 intelligent interactive mobile robot architecture solution, which helps intelligent robot researchers to research, design, development and test intelligent robot efficiently. By configuring complete sensors and actuators, and developing ROS (Robot Operating System) based software system, LAMDA-ROB-2 is able to fully perceive the environment, respond to commands accurately, furthermore, it contains lots of analytical tools to improve the efficiency of development.Secondly, regarding the problem of limited application for single sensing technology, this paper proposes LAMDA-ROB-2 natural human-robot interaction system solution, which combines the new deliverables of different sensing technologies effectively and extends interaction functions easily. By normalizing the input sensing commands, multiple sensing technologies can give the same semantic output; by the routing processing of the semantic commands, the system can make interaction modules independent of each other, thus can be easily expanded. Such approach has been realized in LAMDA-ROB-2, the result has proved that different sensing technologies can compensate each other, and can be expanded without too much effort.Thirdly, regarding the problem of lacking natural human-robot interaction function in autonomous navigation system, this paper proposes LAMDA-ROB-2 autonomous navigation system solution, which makes human controlling robot navigation process with semantic commands possible. By semantically labeling the places in the map, the semantic commands in the natural human-robot interaction are able to match the places in the map of autonomous navigation system. Such approach has been realized in LAMDA-ROB-2, the result has proved that human can control the robot to construct map and navigate with voice and other natural manners.