< Qiong Wang homepage

Qiong Wang

Master Candidate of Science
GRASP Laboratory
Computer and Information Science
University of Pennsylvania
qiong@seas.upenn.edu


Project Overview


Graduate

Vision-based Position and Velocity Estimation with Controlling of Quadrotor more

Vicon-based Path Planning and Controlling of Quadrotor more

Image Rectification and Stitching with OpenCV more

Logo Impainting with Hough Transform more

Robockey: Autonomous Hockey-playing Robot Competition more

Self-balancing Robot more


Undergraduate

Au-Nano Structure Fabrication by the application of Electro-decomposition and Surface-enhanced Raman Scattering (SERS) Detection for Au-Nano Particle more






Graduate

Vision-based Position and Velocity Estimation with Controlling of Quadrotor

March 2013 - April 2013

This is one project of MEAM 620 Advanced Robotics. It aims to estimate the position and velocity based on the image information obtained by the camera. The position was estimated by the data package from April Tags with the determination of the extrinsic camera parameter. The velocity was estimated through the computation of optical flow between two continuous images. After the Extended Kalman Filter (EKF), the actual position was used to control the quadrotor to fly according to pre-set path points.






Vicon-based Path Planning and Controlling of Quadrotor

January 2013 - March 2013

This is another project of MEAM 620. The goal is to plan a 3D path with Dijkstra and A-star algorithm with the controlling of both a quadrotor in simulation and an actual quadrotor. Linear-quadratic Regulator (LQR) method and proportional-derivative (PD) method have been implemented to control the quadrotor. 





Image Rectification and Stitching with OpenCV

March 2013 - April 2013

The project aims to combine multiple photographic images with overlapping areas of view to produce a panorama image. Vanishing points and vanishing lines method and matching points method were implemented to stitch images. Rectification of images were used to find the vanishing points and lines so as to calculate the homography matrix.





Logo Impainting with Hough Tranform

April 2013

As the final of CIS 580 Machine Perception, the objective is to impaint a circular logo into the detected elliptical area with alignment to the vertical vanishing point. Hough transform was used to detect the ellipse from the edges in the original image. The circular logo of BMW was then warped into the elliptical area after estimation of the projection from the detected ellipse to a circle. After finding the vanishing point, the logo was aligned with the rectangle with specific rotation.







Robockey: Autonomous Hockey-playing Robot Competition 

November 2012 - December 2012

2012 Robockey is the final project of MEAM 510 Design of Mechatronic System. Our team built three fully-autonomous hockey-playing robots in five weeks with great hard work to compute in the tournament against sixteen other groups. We were victoriously finishing in the Top 4! The robots were equipped with IR blob detector to find the IR-emitting puck and were able to determine the relative position and orientation on the rink through the camera in the Wii sensor. 



Self-balancing Robot

October 2012 - November 2012 

This two wheel self-balancing robot was built under two-week time by a group of three students. Through only two co-axially mounted wheels touching the  ground, the robot was able to balance itself with robustness though its center of gravity was above the wheel axis. Complimentary filter was used in gyro and accelerator data for proportional-derivative (PD) control of the angle position. Wireless data transmission was implemented for tuning parameters of PD.






Undergraduate

Au-Nano Structure Fabrication by the application of Electro-decomposition and Surface-enhanced Raman Scattering (SERS) Detection for Au-Nano Particle

January 2012 - June 2012 

It is my senior design project. A general SERS detection method was put forward through one-step fabrication of nano composite structure of the substrate and analyte based on the dielectrophoresis and electrodeposition theories. The goal is to meet the needs of extremely high sensitive biochemical detection with easy operation and low costCompared with commonly-used stepwise detection, this method can effectively increase detection sensitivity. The final paper was collected in the Outstanding Senior Design Papers of Tsinghua University.