Game app for pick and place 3D printed robotic arm

This project proposed to build up a “Compact AlphaGo”. For the software part, an Android chess game application was developed by Flutter with a Flask image processing server which was developed by Python. The Minimax Algorithm and Alpha-Beta Pruning constructed a “brain” to the AI which is embedded inside the Android app. A robotic arm with a chess stacker was built by the 3D modeling software AutoCAD, the cheeses were grasped by the robotic arm precisely and placed on the correct grid of the chessboard.

Game Design

In view of the high number of COVID-19 infections in Hong Kong, the HKSAR Government has implemented a social gather ban since 2020. In this game (COVID RUN), a player plays as a HK citizen. He or she needs to escape from the crowd by running in order to survive in the pandemic.

Supermarket Jump!

Supermarket Jump is a minigame where player can take part in a casual parkour experience. The player will play as a box of milk who does not want to be purchased and tried to jump against the conveyer belt to avoid being bought. The main objective of the game is to stay on the conveyer belt as long as possible. The player will encounter various obstacles in the process and some obstacles are deadly – touching them will immediately lead to the end of the game. Bonus points will also be offered if the player successfully performs a “risky jump”.

Soulnos: The Lost Soul

Soulnos – The Lost Soul is a 2D single-player platform game, which player controls two characters to clear every stage. In game, player will face various challenges, including two kinds of monsters, and four kinds of traps. As the movements of the two characters are synchronised, player needs to control the characters smartly to avoid dangers in every stage. Both characters can make a melee attack to kill the monsters in the game. In addition, boosters may exist in each stage, assisting the player to clear the stage. Four kinds of booster, including jump boost, speed boost, healing boost and ranged attack boost are included. When the game is progressing, hints will be revealed to player and player will know more about the truth.

Ditto Flee (PC Game)

The project is a multiplayer PC game that is inspired by hide-and-seek games from the real world. The gaming background is set in a laboratory with the Dittos being caught by the scientists. Half of the players will act as the Dittos to explore the lab and find out the buttons for opening the lab door for escape. Another half of the players will act as the guards of the laboratory to catch the Dittos preventing them from escaping.

3D Open-World Game

Harae – The Legend of Haruto”” is an open-world, Action RPG game inspired by games such as “”TES V: Skyrim””. The story is set in a post-apocalyptic world where robots have taken over humanity by unleashing a deadly plague onto the world. Haruto, the protagonist, seeks to save human-kind by destroying the robots’ source of power, located on his birth-island.
The goal of this project was to create as lively, detailed, and performant a game as was possible in the short development time that was given for COMP3329, and it will continue to see updates even after the semester is over.


This 3D platformer running game is about a girl falling into a dream which guides her into a world called Roll-topia. In the communication with the residents of the local villages, she learned that the villagers were frustrated with the invaders who destroyed their houses. As a result, she decided to kick up the invaders. She uses cheese to attract attention to the greedy mice on the way to protect herself. In the end, she surprisingly arrived to the beginning point, only to find out this world is in the shape of a hamster wheel.

Smart Soccer Robot (ii)@4th Inno Show

This is a remote-controlled robot designed to play soccer matches against other robots. It is a product aimed at teenagers learning STEM education. Our design is inspired by the shape of a charging bull, which makes it more attractive to our target audience. The robot is made from mBot parts, laser-cut wooden panels, and 3D printed wheels. It is assembled using only a few screws due to our use of interlocking mechanisms integrated into the panels. The robot includes an optical sensor, and displays the number of times it crosses a black line on the seven segment display on the back.

Smart Soccer Robot (i)@4th Inno Show

Our product is a smart soccer robot that is primarily used to revolutionise the technical education industry. We designed it on SOLIDWORKS, being very careful regarding technicalities like the position of the centre of gravity and weight distribution. The materials are all environmentally friendly so that we fulfil our ethical responsibility of keeping the environment clean. The robot’s design allows it to “engulf” the ball, keeping it secure for as long as necessary. The precise mBlock programming allows it to be manoeuvred professionally. Both of these features, when coupled together, make it quite a useful product in terms of efficiency. In addition to dominating on the field, it can also serve purposes like spreading the influence of STEM education by making it more fun and stimulate students to work in teams and think critically.

Knowledge Graph Database for COVID-19 Drug Discovery

This project uses knowledge graph database analysis for repurposing existing drugs for COVID-19 treatment. The knowledge graph used in this projected was populated with various kinds of bioinformatic data relevant to the COVID-19 disease, the SARS-CoV-2 virus that causes it, as well as several thousand drugs. This data was then analyzed using three prevalent knowledge graph link prediction techniques: motif-based analysis, PageRank analysis, and machine learning analysis. This was used to generate scores for each drug in the database, with a higher score indicating greater probability in treatment of COVID-19. These proposed drug rankings were then cross-referenced with data on ongoing and completed COVID-19 drug trials, and it was found that of the top 20% proposed rankings, 120 are already undergoing trials, indicating high prediction accuracy. Additionally, as the main users of the project are supposed to be medical experts with non-technical backgrounds, a clean and easy to use interface was also designed for this system. The user interface consists of a proposed drug rankings table, a visualization of the scores of the top 50 drug candidates in the form of a bar graph, and list of literature relevant to each proposed drug along with sentiment analysis of the author’s attitudes towards use of each drug towards COVID-19 treatment. These features enable medical experts to identify the most promising drug candidates so that they can quickly move to the drug trial stage.