Projects

Course Projects


Project Course

* We explored how automated reasoning tools could be used to enable robot programmers to build state machines that can be extended at run time based on user interaction. Specifically, we provided a framework in which programmers can axiomatize their states in terms of their requirements and effects on the state of the world, then leverage this logical representation of the states to dynamically plan sequences of actions to accomplish unforeseen goals (python, Z3, ROS)

CSE507 Computer-Aided Reasoning for Software Engineering (UW)

* Designed an orbit transformation of cubesat towards an asteroid using low-thrust techniques

SD2920 System Integration for Space Technology, Part 1 (KTH)

Collaborators: Shuqi Xu and Mikael Andblom
* Learned how to model and solve problems by using GAMS;
* implemented the active-set method, the interior-point method for QP and the SQP (Matlab);
* implemented Projection Onto Convex Sets (POCS) and other methods based on the fixed point theory of class of nonexpansive mappings (Matlab);
* implemented Nonlinear MPC with SQP solver (Matlab)

SF2822 Applied Nonlinear Optimization (KTH)

Collaborators: Sylvain Potuaud and Beatrice Ionascu
* Coded self-organizing map, Hopfield, etc. (Matlab)

DD2432 Artificial Neural Networks and Other Learning Systems (KTH)

Collaborators: Shuqi Xu
* Proposed a hierarchical structure to multi-robots so that they can map and localize themselves efficiently and robustly;
* implemented the proposed EKF-based multi-robots SLAM by using the skeleton code offered by the course (Matlab)

EL2320 Applied Estimation (KTH)

Collaborators: Arvid Fahlström Myrman, Axel Riese, and Peter Langenberg
* Coded group factor analysis from scratch (Python)

DD2434 Machine Learning, Advanced Course (KTH)

Collaborators: Robin Maillot
* Built a robot that tracks colored circles using Arduino and Raspberry Pi (C, Python);
* made the robot detect human faces by using OpenCV;
* made Raspberry Pi send emails, receive commands via twitter, and fetch temperature information from websites (Python)

EL2222 Systems and Control in Practice (KTH)

* Coded deep learning from scratch, including back propagation, batch normalization, drop out, weight regularization and stochastic gradient decent (Matlab);
* Coded a recurrent neural network and trained it to learn excerpts from a novel and tweets

DD2424 Deep Learning in Data Science (KTH)

Skills

Languages


Japanese: Mother tongue
English: Advanced (March,2019 TOEFL iBT 107/120)
Chinese: Intermediate (Started learning in 2010)
Swedish: Beginner (Passed B1/B2 level at KTH)