Projects

Partnered with Boeing to explore single pilot cockpit operations in commercial aircrafts.

Overview

Designed and developed a proof-of-concept solution for a single pilot cockpit using smart glasses and an automated cockpit switch system. The solution handles tasks typically performed by a co-pilot, such as executing checklists, error detection, and ensuring pilot alertness.

Contributions

1. Engaged with Boeing engineers and pilots to understand current cockpit procedures
2. Designed functionality of smart glasses and automated switches:
• Overlays critical information such as checklist items and warnings in the pilot's field of view
• Tracks the pilot's blinking pattern and wakes them with vibrations if their eyes remain closed for an extended period
• Uses computer vision to detect errors in pilot's actions and alert them accordingly
• Automated switches execute tasks upon pilot approval while maintaining manual override capability as backup
3. Sourced and integrated all electronic components, including a mini OLED display, vibration motor, motion sensor, video camera, buttons & switches, rotary motors, and microcontrollers
4. Developed all functionality of smart glasses and automated switches including microcontroller algorithm, computer vision, and OLED display content (Yolov8, Python, Arduino)

Results

The solution offloads lower-level tasks, such as manually flipping switches and looking down at cockpit instruments. This helps the pilot to focus on higher level tasks like decision making and monitoring the flight path, thus improving the pilot's situational awareness and flight safety.

NUS Computer Engineering course on Real-Time Operating Systems.

Overview

In this project, teams are tasked with building a remote-controlled robotic car capable of navigating an obstacle course as quickly as possible. Each car is built using similar components, with the primary differentiating factor being the software and operating system developed by each team. The cars require essential functionalities, including playing music, flashing lights, and maneuvering abilities. The car uses an ESP32 for wireless control and a main microcontroller (FRDM KL25Z), which contains an ARM Cortex M0+ processor. This processor runs a Real-Time Operating System (Keil RTX), programmed with threads and interrupts in C.

Contributions

1. Co-designed the OS architecture
2. Recommended switching from HTTP to UDP protocol for control signals, reducing latency
3. Implemented the flashing light functionality requirement
4. Designed and procured a custom laser-cut acrylic chassis frame to enhance vehicle dynamics
5. Drove the car during final obstacle course run to obtain the team's timing

Results

The thoughtful design and implementation of our Real-Time Operating System led to exceptionally low control latency. Additionally, the joystick control system, developed by a team member, provided finer control over the car's movements compared to traditional button controls. This responsiveness and ease of control enabled our team to achieve the fastest lap time in our lab class by a comfortable margin.