Abstract : To solve the problem of autonomous control of spherical robots, an experimental platform for spherical robots based on monocular vision and ultrasound was built. RGB threshold segmentation based on local images was used to identify target objects and calculate the center coordinates of the target objects. Distance information was obtained through an ultrasonic sensor. Control rules were designed by integrating information from multiple sensors, and target tracking was achieved through servo testing of the robot. Keywords : Visual tracking; Multi-sensor; Image segmentation; Spherical robot 1. Introduction Autonomous mobile robots are a hot research area in robotics, involving many aspects such as actuator control, sensor data fusion, image processing, pattern recognition, and neural networks. Spherical robots, as a type of mobile robot, have omnidirectional motion characteristics. They can achieve both the fast walking performance and high load-bearing capacity of wheeled robots, and the adaptability to uneven terrain like walking robots. Therefore, research on the autonomous control of spherical robots has high application value. The experimental objectives of the spherical robot studied in this paper are as follows: First, the target object is placed in the test area, but not within the robot's visual range. The robot circles in place to find the target object, and drives the target object's orientation based on the set image features of the target object. Then, it approaches the target object and maintains a specific distance. Next, the test personnel move the target object, and the robot tracks the target object's movement. 2. Hardware System Platform Description The spherical robot used in the experiment is the BYQ-8 spherical robot (Figure 1) developed by Beijing University of Posts and Telecommunications. The robot's outer shell consists of two independent hemispheres, each driven by a separate left and right motor. The short-axis motor controls the lateral tilting or uprighting. [Full text download of the research title "Research on Multi-Sensor Target Tracking of Spherical Robots"]