Arduino with mini-robot car

In this project, we used Arduino to control a mini-robot car following the tracks below.

Case	Left sensor	Right sensor	Bumper sensor	Left motor rotation	Right motor rotation	PWM	Car movement
1	1V	1V	1V	forward	forward	140	forward
2	0V	1V	1V	backward	forward	100	Turn left
3	1V	0V	1V	forward	backward	100	Turn right
4	0V	0V	0V	Backward	Backward	140	backward

Smart Door Locking System

In this project, we used Keil and CubeMx to implement a smart door locking system using STM32. The motor represented the door. If user typed in correct password through the keypad, the door opens and the green lights on, else, the door remained close and red light on. If 3 consecutive wrong inputs were detected, the buzzer produced noise for warning and sent alert message to the phone of owner through bluetooth.

CubeMx Configuration

⬩ GPIO

⬩ PB0, PB1, PB5 -> RGB LED

⬩ PC0, PC1, PE5, PE6, PA4, PA5, PA6 -> keypad

⬩ RXD(PA9), TXD(PA10) -> Bluetooth module HC-05

⬩ PB6 -> motor

⬩ [reserve PC8 to PC12 for SDIO]

⬩ Enable Real Time Clock (RTC) -> get the time

⬩ Enable FSMC -> LCD

⬩ Enable TIMER-> rotation of the motor

⬩ Use TIM3 and TIM4 to generate PWM

⬩ Pulse = 900__close the door, Pulse = 300__open the door

⬩ Enable USART1 -> make use of the HC-05

⬩ Enable SDIO -> make use of the Micro SD card

Monster Matrix Game

In this project, we used the concept of linked list from C++ to implement a monster killing game.

Upper case characters represent different type of monsters

Lower case characters represent different type of weapons

Integer number represents the number of things(monster / weapon)

Different monster can only be killed by specific weapon

Kill all the monster to win the game

Chatroom implementation

In this project, we used C language to perform a socket programming to implement a chatroom, a communication plateform for multiple users.

Support multiple clients chat at the same time

Support broadcast message and private message

Each user has a unique name, Same name error checking

Final Year Project

● use python to build the GAN algorithm model for image deblurring on Jupyter Notebook

● train and test our model with GOPRO dataset

● convert kernel functions, zero padding & 2D Convolutional layer, from python codes into C codes

● pipelined them in Vitis HLS for simulation

● achieved an acceleration of the performance by more than 10 times.

OTHERS