An Object Localization System Using Monocular Camera and Two-Axis-Controlled Laser Ranging Sensor for Mobile Robot

Abstract

Aiming at object grasping of indoor mobile robots, this paper develops an object localization system based on the fusion of a monocular camera and a two-axis-controlled laser ranging sensor. The system can locate the specified object in space without obtaining the background depth information of the object, which reduces the consumption of computational resources and has a low cost. According to the coordinate mapping of the object in the image, the measurement point of the laser ranging sensor is controlled to coincide with the target point by two servo motors in perpendicular axes. Thus, the spatial location of the target point can be determined from the rotation angles of the cradle head motors and the laser-measured distance together. To accurately control the rotation angles, a compound strategy combining feedback and feedforward is proposed. On the one hand, the feedback of the laser measurement point is introduced in the image coordinate to realize the close-loop control of the measurement point. On the other hand, a feedforward compensator based on disturbance observation is designed to reduce the influence of robot motion. An experimental setup is developed and substantial experiments are implemented to evaluate the performance of the proposed system. The results show that object localization can be effectively achieved with good accuracy and dynamic response.

Keywords

• Mobile robot
• object localization
• sensor fusion
• laser ranging sensor
• disturbance observation