Planetary Rover Manipulation

A project of the Robotics 2019 class of the School of Information Science and Technology (SIST) of ShanghaiTech University. Course Instructor: Prof. Sören Schwertfeger.

Liang Chen,Ruiqing Chen,Yang Li

Introduction

Nowadays people are basically superseded by robot in performing risky and dangerous task which are not to done due to size, environment constraints.  One of the most important part in this is manipulation, which are undergoing important and dramatic changes daily.  And advancements in the field of automation and robotics technologies are also contributions towards it. The project of the Robotics is rover manipulation, which need the robot find where is the shovel firstly, then move the robotic arm to close to the shovel, grasp the shovel and then move it to the area of scoop sand. Having done this, the robot needs to move the sand carefully to the specific place. Machine vision has better sensitivity and resolution.  It tends to application for identifying and inspecting details, guiding the robot controllers.   Vision is very important in automation industry, which can be used for number of industry application such as welding, painting, automatic pick and place object, military operation, in space, in underwater operation, in agriculture operation1

etc.  One of the function of vision is to identify shovel, which the robot know where it is and how to proceed in order. Specifically, we will fix the ZED Camera on the side of work scene to obtain the height of sand, which based on pointcloud from ZED camera,  the ZED point cloud was used algorithm to get the height of sand.  Having done this, the robot can move the arm to the place of shovel and move it to do some tasks. 

Experiment

The test tasks will be divided into two steps:

•Experiment 1, the robotic arm holds the shovel and knows where the sand is, and then uses an algorithm to move the robot arm to the sand, shove the sand, and finally pours the sand into the pre-prepared collection device.

•Experiment 2, the shovel is somewhere on the ground, the robotic know its position, then move there, and grab the shovel, move to the position of sand to shove the sand, but robot don’t know the height of sand, so we need depth camera to get it, and finally the sand in the shovel is poured into the pre-prepared collection container.

In this project, a ZED camera will be used to get the height of sand.  Then using ROS packages to drive the robot arm to grasp shovel and sands.  Having done this, the robot smoothly transports the sands to the specific place or someplace the robot don’t know, the position of destination can be obtained by depth camera.

 

Conclusion

In this project, we use RANSAC segmentation algorithm to calculate the height of sands, and also use ros package kinova, pcl and moveit to process information and proceed task. We have met some trouble when proceed the motion plan, but finally we find method to solve it and accomplish this task successfully. This project still has some problem need to improve, for instance, use laser or camera to locate the position of object automatically, find more efficient ways to decrease the probability of obtain wrong motion planning trajectory and how to design motion to grasp more sand. Maybe it will be improved in the future.