Yunxiang Yang(杨云翔)

PhD student at the University of Georgia, Athens, GA.
Email / Website / LinkedIn / GitHub

🎓 Education

University of Georgia Athens, Georgia (Jan. 2023 - Present)

• School of Environmental, Civil, Agricultural and Mechanical Engineering
• Research Interests: Time-series traffic data/Deep Neural Netweoks/Civil Applications/Signal Processing

Stevens Institute of Technology Hoboken, New Jersey (Jan. 2020 - May 2021)

• School of Electrical and Computer Engineering
• M.S. in Electric Engineering with an AI concentration
• GPA: 3.934

Changchun University of Science and Technology Changchun, China (Sept. 2014 - June 2018)

• School of Photo-Electronic Engineering
• B.Eng. in Engineering

🧑🏻‍💻 Work Experience

United Parcel Service (UPS) Mahwah, New Jersey (June 2021 - Feb. 2022)
Position: Data Scientist in Advanced Analytics Group(AAG)

• Address proximity:
• Built delivery frequency, CR Ratio, geographic coverage, and other useful features to reveal important characteristics of Illinois Jefferson Hub and laid a good foundation for later research
• Solved customer similarity, defined all the unit areas into 3 classes, and employed KNN to determine the optimal parameters in ST-DBSCAN. Eventually reduced the amount of calculation by 75.9%
• Customized ST-DBSCAN model into “frame” mode (iterate adjacent unit areas), fundamentally avoiding memory crashed issue and significantly reducing the training time consumption by 46%
  HR Peak Season Hiring Plan:
• Extracted and organized the HR hiring plan data for all the UPS facilities in the U.S. region
• Completed the data engineering part and made the data accessible on Google Cloud Storage
• Finished the payroll analysis and explored the employees’ characteristics of each facility
  Aggressive Driver Behaviors Analysis:
• Built and trained neural network model LSTM on Google Cloud Platform (GCP) with public time-series sensor data
• Implemented Random Forest Model to solve class imbalance and improve the overall accuracy to 91%
• Returned the trained model to the Flask server then displayed with a web-based dashboard

Stevens Institute of Technology Hoboken, New Jersey (Feb. 2021)
Position: Grader of EE575 Control Systems

• Graded assignments to provide timely feedback for graduate students

🧑🏻‍🔬 Research Programs

2021 Master Thesis in Electrical Engineering Hoboken, New Jersey (Feb. 2021 - May 2021)
Objective: Contactless vital signs measurement and analysis via FMCW radar
Laboratory: The Bio-Electromagnetics Laboratory at Stevens Institute of Technology
Advisor: Associate Prof. Negar Tavassolian Google Scholar
Reader: IEEE Fellow Prof. Lei Wu Google Scholar

• Application of implementing Ultra-Wide Bandwidth Frequency-Modulated Continuous Wave (UWB-FMCW) radar to detect human vital signs.
• The wearable sensor shimmer3 GSR + and Apple Watch 3 were used as reference signals.
• The phase extraction region is determined by range FFT, and then the phase is extracted by vibration FFT. Through DC removal, Hanning window and two band-pass filters, human respiration signal and heartbeat signal were extracted respectively.
• After calculation and Pearson correlation coefficient (PCC) analysis, 96.4% correlation (Apple watch 3 as reference signal) and 95.5% and 83.8% correlation (shimmer3 GSR + as reference signal) were achieved respectively.

2020 ECE Student Honor Summer Research Hoboken, New Jersey (May 2020 - July 2020)
Objective: Implementing and accelerating Transformer and BERT by using CUDA C/C++
Laboratory: HPDA (High-Performance Data Analytics) Lab at Stevens Institute of Technology
Supervisor: Associate Prof. Hang Liu Google Scholar
Position: Leader of a 3-person teamwork

• Most of the work was done on the XSEDE cloud server which has a huge amount of GPU resources.
• Used Shared Memory & Memory Coalesced methods to optimize matrix multiplication and matrix transpose.
• The time to complete an epoch with shared memory optimization and non-optimization are 1.506s and 1.787s respectively.

Undergraduate Innovation and Entrepreneurship Training Program Changchun, China (Apr. 2016 - Apr. 2017)
Objective: To design a non-contact palmar vein recognition instrument based on image processing
Supervisor: Associate Prof. Shifeng Wang
Position: Leader of a 3-person teamwork

• Tried the methods of open-close operation, Gaussian filtering, NiBlack local binarization, adaptive threshold, dilation, and erosion to perfectly show palmar vein image.
• Used the ORB test method to extract the key points of two images that need to be matched.
• Utilized a mathematical method RANSAC for eliminating mismatched points.

Study of Vortex Beams Based on LCOS-SLM Changchun, China (Apr. 2016 - Apr. 2017)
Objective: To implement vortex beams based on LCOS-SLM, then propose a new method producing vector vortex beams based on LCOS-SLM.
Supervisor: Prof. Zhiying Liu
Position: Leader of a 3-person team work, but the only one sticked to the end.

• Adopted computer numerical simulation to compute the interference of object beam and the reference beam.
• Got vortex beam and optimized index properties by Combining the algorithm principle of CGH and LCOS-SLM.
• Got an Interferometric hologram.

📌 ACADEMIC CONTESTS

The 2nd National “Intelligent Internet Competition” The Second Prize Chongqing, China (Mar. 2017 - Aug. 2017)
Objective: Designed an intelligent non-contact heart rate measurement, guaranteeing the baby is left in a safe unattended environment.
Position: The ONLY ONE out of 589 nationwide teams who are sponsored by ourselves

• Developed facial recognition and identification based on OpenCV.
• Processed the images of the forehead area and crawled the signals of heart rate.
• Used STM32 to extract sensor signals, and controlled the relevant appliances by monitoring heart rate, such as adjusting fan speeds (fan is used to reduce the concentration of carbon dioxide).

2016 National College Students Optoelectronic Design Competition Chengdu, China (Mar. 2016 - July 2016)
Objective: To design an autonomous airdrop UAV based on monocular vision.

• Developed facial recognition and identification based on OpenCV.
• Implemented the method of GPU parallel computation to process the images.
• Established an embedded system including NVIDIA Jetson TK1, Arduino, and APM.

Graduate Academic English Speech Contest The Second Prize Changchun, China (Nov. 2015 - Jan. 2016)
Made a speech in English, expressing personal opinions on image processing based on OpenCv, the understanding of the SLAM system, and its application in automatic driving fields.

The 5th OptoElectronic Design Competition The First Prize Changchun, China (Apr. 2015 - Oct. 2015)

• Designed an optoelectronic intelligent toy car based on HSV color recognition.
• Applied an Ant colony algorithm to the car, which made it possible to pick up 10 ping-pong balls within 35 seconds on a 4-square-meter ground.
• Applied dynamic PID control method

🎤 Public Speaking

2021 UPS Summer Internship Presentation Intern representative Mahwah, New Jersey (Aug. 2021)

• Selected by IT Vice President Ken Finnerty and Senior Director Mallory Freeman.
• Made a presentation to UPS CIEO Juan Perez on behalf of all interns in the IT department.
• Talked about internship project address proximity and the business values behind it.

Interviewed by China Youth Daily Changchun, China (June 2017)
As the lead director of Yullion Studio got interviewed by China Youth Daily, and the interview report got reprinted by China Education Network, People’s Daily Online, and other mainstream media.

Photo-Electronic Design Contest Experience Sharing Meeting Keynote Speaker Changchun, China (Nov. 2015)
As the keynote speaker, hosted and organized a sharing meeting with more than 100 people. At the meeting, we shared the technology and knowledge used by our team to win the photoelectric design competition, so as to help students broaden their horizons.

🏆 HONORS AND AWARDS

• 2021.7 UPS Hackathon Final FAB FOUR (4/24), presented to CEO&CIEO at UPS.
• 2020.7 Third-prize in Stevens ECE Honor Summer Research Program, University Level
• 2018.7 Excellent Graduation Design, University Level (10/459)
• 2017-2018 Third-prize Scholarship, University Level
• 2017 Certificate of “National Level of Electronic Information Professionals” by Chinese Institute of Electronics, Assistant Engineer Level
• 2017 Second-prize in National Final of Intelligent Internet Competition, National Level
• 2016-2017 Second-prize Scholarship, University Level
• 2015 Second prize in Academic English Speech Contest, University Level
• 2015 First prize in Photo-Electronic Design Contest, University Level
• 2014 Third-prize Scholarship, University Level

💡 Skills

• Computer Vision(OpenCV), Hyper Performance Computing(CUDA in NLP)
• Windows & Linux Heterogeneous Programming(C++), Python, SQL, BigQuery, Matlab
• Google Cloud Platform, InfluxDB Cloud, Grafana
• Raspberry Pi, NVIDIA Jetson TK1
• Arduino, AVR, STC51, Quadrotor UAV, APM
• UWB-FMCW Radar, Hukoyo Single Line LIDAR, Velodyne 32 Lines LIDAR