(A) study on the AI-based navigation and obstacle detection for mobile robots

Abstract

For mobile robots to be autonomous, they should have essential functional capabilities such as real time obstacle range sensing, determination of their current locations and heading angles, local path planning for unknown environments. In order to endow the mobile robots with these capabilities, this thesis deals with the above three issues. For real time obstacle detection, multiple ultrasonic sensors mounted on the front face of mobile robot are used. These ultrasonic sensors can be periodically calibrated using the rotating ultrasonic sensor mounted on the top of pari-tilt device. In order to avoid the crosstalk between multiple sonar sensors, the scheduled firing method is presented. The locations of obstacles are registered in the local map defined at the mobile robot frame. The local map can be constructed and updated in real time. Generally, mobile robot has to know its current location and heading angle by itself as accurately as possible to successfully navigate in real environments. To achieve this capability, we developed a mobile robot localization system which uses two cylindrical beacons and a single rotating ultrasonic sensor. The proposed method can estimate the position and heading angle of a mobile robot using the sonar scan data obtained at a single mobile robot location. To acquire the center positions of two beacons from the sonar scan data, a data processing algorithm was developed. Using this algorithm, we could accurately obtain the geometric parameter sets of beacons and, consequently, determine the position and heading angle of the mobile robot. To show the effectiveness of the proposed method, the experimental results of the proposed method were compared with those of the dead reckoning method in the presence of wheel slippage. For the local path planning for unknown environments, we proposed a Al-based navigation method utilizing the fuzzy logic and reinforcement learning method. The proposed navigator consists of avoidance and goal-...