Electrical Engineering graduate researcher

YI PENG

**** ******* ***, *** *** SAN DIEGO, CA, 92122 310-***-****

adckhg@r.postjobfree.com

EDUCATION

University of California, Los Angeles

Bachelor of Science, Electrical Engineering, Technical breadth in Computer Science, Cumulative GPA 3.45 University of California, San Diego

Master of Science, Electrical and Computer Engineering, expected June 2020, GPA 3.57 RESEARCH

Graduate Researcher Summer 2019 - now

UCSD Advanced Robotics and Controls Lab, Dr. Michael Yip

Data-driven selection and design of muscle-actuated biomimetic motions. Utilizing machine learning algorithms such as EM for finding missing data and modeling Gaussian mixtures, SVM and K-means.

Developing a generalized modeling and control strategy common across robot muscle groups. Programming neural networks, designing open- loop control systems.

Building robot arm with 3D printing to test out actuator model. Technical Projects

Reinforcement Learning Algorithm Design Fall 2019

Applied divide and conquer policy using k-means classification to reinforcement learning algorithms such as TD3 and TRPO.

Tested and compared results on OpenAi Gym environments Ant and Half Cheetah, significant increase in returns. Multi-Agent Actor-critic reinforcement learning methods Fall 2019

Implemented A2C algorithm in mine-field game, where multiple agents navigate to a single flag goal.

Use CUDA kernel and parallel programming to accelerate updates on each agent during iterations. RoboCoach-Healthcare Robot Project Spring 2019

Designed robotic health coach assistant using TurtleBot.

Robot outputs display and speech to give user directions. User able to interact with robot using keyboard inputs.

The robot can act as a walking companion. Uses SLAM Map building to lead user to specified 4 destinations.

Interactive exercises use color recognition and pattern recognition to determine if the user is doing the exercise actions correctly. The robot also gives advice to user if actions are not up to standard. Moderator Robot- Human Robot Interaction Design Project Winter 2019

Designed robot to act as Moderator in Mafia game.

Hand gesture recognition via hand recognition and classification using convolutional neural network.

Robot motion control using ROS indigo.

Interface developed using pygame. Uses TCP sockets to communicate with hand gesture recognition and motion control.

Pencil Sketch to Photo Realistic Portrait Fall 2018

Generate a photo-realistic picture based on the corresponding pencil sketch, using implementation of conditional version of generative adversarial nets (CGAN).

SKILLS

Programming in C++, C#, Python, Matlab, JavaScript, Clisp, CUDA GPU parallel programming, swift.

Familiar with pytorch and TensorFlow in deep learning applications.

Computer vision applications such as color and shape recognition, feature detection, camera calibration.

Computer graphics and animation with OpenGL.

Comfortable with path finding algorithms, reinforcement learning algorithms, and SLAM.

Unity engine in development of VR games, animation, and motion capture.

Contact this candidate