Github
LinkedIn
EmailSelected Projects
Tiling Robot Design
In this project, I participated in a tiling robot for room and hall. I am mainly responsible for Aubo Force Feedback System and Mortar Injection process. To build Force Feedback System, I attempted to resort to torque sensor but the weight of sensor seriously affect the working efficiency of tiling. So, a simpler but reliable method is necessary.
1. Aubo Force Feedback System
I visited Aubo Offcial SDK and collected electrical current of every joint. From a paper , there should be a linear regression model for currents and joint torques. In this way, I used torque sensor to collect torques and corresponding currents. After that, I trained a linear regress model to predict torques by currents.
Then, based on equation ,
I used MoveIt! to obtain the joint states and Jacobian
matrix based upon URDF of Aubo Robot to compute the force and torque in the end-effector with the help of
Eigen
in C++ for matrix evaluation. The computed force and torque in the end-effector are published in a /Wrench
topic
as WrenchStamp msg with a frequency of 100 hz.
Then, I put forward a force feedback system. The principle is that the laser in the end-effector will keep inspecting the distance between the end-effector and object and reminds system to mark current force as force threshold when the distance becomes smaller than the distance threshold which is set according to experience in various working environments. Then, as the end-effector constantly closes to object, the system will examine whether the force difference between current force and force threshold is larger than the set force_difference_threshold. Once the system confirms the result that force difference is too large, the state of robot will be updated as touched and follow up a series of operations according to the robot task.
2. Mortar Injection
After completion of Force Feedback System, I am assigned a project to finish the moratr Injection process design. In my design, the robot arm will initially to move to a position to scan the ARcode by camera to recoginize the accurate position of the board. Then the robot will plan to suck the board with the help of the Aubo Force Feedback System. The detail of operation is shown in the video.
Once sucking the board, the robot arm with board will move to initial position to start mortar Injection. Based on reacheable range of robot arm and the reason that we are not willing to inject the mortar on the surface of tile to lower the cleanness, the robot will autonomously adjust the injecting range (with and length) according to the position of corner of tile of last unit in the base link coordinate, which will be completed before sucking board and is a operation that the robot arm will move to neighbor of tile of last unit to compute the corner position. Then, I implement computeCartesianPath in MoveIt! to finish the injecting process as the video shows.
The final result could be like the picture below with the help of scraper.
This is our group photo.
"
IoT Project: Smart Movable Trash Bin
Smart Movable Trash-Bin (SMTB) is an IoT solution to current trash sorting problem in office. Currently, the trash sorting is a trend and a remedial action to save the environment, which becomes a policy of great significance in numerous countries and cities. However, people, especially the officer in office, who are not going to have a single trash-bin suffers from the trash sorting and often feel confused about the trash type every time they are going to drop trash. Our project aims to automate the process in the origin, which means when people drop trash, the trash is accurately identifed as the trash type, including metal, plastic, glass, paper, cardboard (recyclable in most regions) and trash (unrecyclable). In addition, our target is to make it as a intelligent project. Since in most cases, officers should share a common trash-bin and need to move when they have trash. But with our design, user can directly use phone to call our SMTB to come to his/her postion and user can simply put trash in and SMTB intelligently identify the trash type and go back orginal place for sorting and collecting.
With SMTB, officers will benefit from simplification of dropping trash and trash sorting; office will become more clean; government will benefit from the energy conservation resulting from resource reuse. We hope our project will make a difference to the world.
The details of this project is shown in the page IOT SMART MOVABLE TRASH BIN PROJECT and the code in my github repository.