| With the development of scientific and the advancement in robot technology, theperformance and intelligent of robot have been constantly improved, has been widely used inthe life of service, military, industrial production and exploration of unknown areas and manyother areas. The autonomous mobile robot is a very important branch in the field of roboticsresearch, the navigation and obstacle avoidance of autonomous mobile robot has always been ahot research field of intelligent robots, among which the robot obstacle avoidance technique isthe key technology to reflect the robot intelligence.what can we do to let the robot know whatare obstacles, and how to avoid them, these two questions correspond to the process of obstacledetection and route planning of the obstacle avoidance. In this paper, research work based onthe autonomous obstacle avoidance problem based on multi-sensor of the indoor mobilerobot.The main content of the research as follows.1. Firstly, we do a brief analysis of the developments in the field of mobile roboticsresearch, the methods of the obstacle detection of the mobile robot, the methods of multi-sensorinformation fusion and the status of the local route planning of mobile robot, and make a briefintroduction of the major component of the research topic.2. Obstacle detection and location based on vision. Firstly analysis the image by theknowledge of information entropy, complete the obstacle detection and preliminarylocation judge by the changes of the value of the information entropy of the image, andlocate the obstacles accurately through analysis the edge points in image.3. Obstacle detection and location based on multi-sensor information fusion. Verify theresults of the visual obstacle detection by the ultrasonic sensor, and integrate the bothinformation get more accurate information about the obstacles.4. The route planning of robot obstacle avoidance based on behavior. This methoddecompose the robot control unit to simplify the path planning of robot obstacle avoidance. Wewill divide the robot’s control unit into to5behavior,include the behavior of find the targetpoint, the behavior of towards the target, the behavior of general obstacle avoidance, thebehavior of the solve to the deadlock, the behavior of the target point is reached. We use themethod of Improved artificial potential field for obstacle avoidance behavior, the behavior module of the solution of deadlock is based on the omni-directionally search solution.5. The robot control based on fuzzy control strategy. When the robot move to the targetpoint or move to one direction, the robot control in decision-making, depending on thesurrounding environment information and practical tests can not make the robot very accuratestraight in one direction,so we can use the fuzzy control strategy to control the robot. Thefuzzy control method in this article is to control the left and right wheels’ speed according to theangle θ,which is the angle of the direct front of the robot and the line between the point of thetarget and the robot. Fuzzy lookup table corresponding to the lessons learned by a large numberof experiments.6. Obstacle avoidance based on artificial potential field algorithm. The basic principle ofthe artificial potential field method is: we always abstract a force field exist in the movingenvironment of the robot, the robot move in a central force field, its direction of motion is theresultant force of all force, see the robot as a particle under the influence of artificial market.The core idea in the artificial potential field method: construct the gravitational field of thearound the obstacles and the repulsion field around the target point, the direction of the motionof the robot determined by the negative gradient direction.7. The improvement of the artificial potential field method. By analyzing the twoquestions exist of the local route planning based on artificial potential field:destinationunreachable and the local minima. Established a new potential field function to solve theproblem of destination unreachable; omni-directionally search solution to solve the localminima problem, the method is the same algorithm as the solution of the deadlock behaviorused in the robot control unit.Robot move forward to the target point,as can be able to identify the target point by thePanoramic camera, and can avoid obstacles and go on the behavior when encounter obstaclesduring the way, and the robot can void obstacle even encounter the local minima problem in theartificial potential field method.The next step is combine the obstacle avoidance method of this paper presents and thenavigation system our laboratory group has completed years before,apply the theoretical studyof the subject application to the practice. |
PDF Full Text Request