Route Planning For NAO Robot

NAO robot is a multi-purpose humanoid robot, which slowly comes into people's lives depending on its outstanding shape and advanced human-computer interaction system. However, NAO robot can move freely like human and obey mankind's instructions to complete various tasks only if it can see the environment around and arrange the route according to the changes of environment, and finally reach the destination and finish task.Combining walking characteristics of NAO robot and sensor configured on it, this article set an effective route planning stratage in order that NAO robot can move freely in complex environment.The core content of route plannig stratage is algorithm, this article adopted ant colony which has inherent advantage in seeking route in complex environment. Ant colony doesn't need too many parameters and computes in short time,the route planned by it has ideal length; meanwhile, ant colony has defalts, it is easily trapped into local optimum and hardly get global optimum solution, it cannot get rid of U trap automatically. This article improved the ant colony to enhence efficiency, introduced ant unnormal die mechanism to get rid of U trap problem, modified parameters to adapt algorithm to route planning for NAO robot.The core content of this article is the design of complete route planning for NAO robot. The main process of route planning is: First, get complex environment through Xtion equipemnt installed in robot, and use raster map generated by octree algorithm as priori map, and set a destination node for the robot, the initial position of robot is given by octree algorithm;Then call ant colony to make the complete route planning and command robot to walk, the sonar equipment in front of robot chest is enabled and monitors the situation in front of robot in real time while it is walking, if the robot encounters moving obstacles, it stops and chooses corresponding measures according to predetermined precept and continues to walk till arriving at destination positionIn terms of NAO robot dynamic.route planning, predetermined precept is particularly important. This article approched a simple and practise method, which is feasible and fast responded to deal with complex environment. In process of dynamic obstacle avoidance, if the robot encounters obstacles it stops, and updates map to replan the route using Xtion, after three seconds it redetects the obstacles ahead, if the obstacles leave, it continues to move according to original route; if the obstacles stay, it avoids obstacles and move on according to the new route.This article tested the program above by arranging real environment and got ideal consequence that the robot completed assigned action according to planned process and arrived at destination.There are still room for improvement such as deep optimization of ant colony and introducing deep learning in order that NAO can be more humanized and walk freely.