🌐 Project Overview: The Philosophy of the "Smart-Rover" Positioning: Not a Toy, but a "Minimum Viable System" This project is positioned not as a toy, but as a "Minimum Viable System." While modern AI—driven by ChatGPT and autonomous driving headlines—is often presented as a magical emergence of data and scale, our goal is to investigate the stark reality. As a machine learning student familiar with OpenCV and TensorFlow, I discovered a profound gap between the digital world of cloud GPUs and the physical world of a handheld vehicle. This gap cannot be crossed simply by tuning parameters. It forced us to confront three fundamental levels of autonomous driving: Perception Layer : Investigating the accuracy gap between basic ultrasonic sensors and industrial Lidar. Decision Layer : Comparing the reliability of classic PID control versus the adaptive "intuition" of Deep Learning . Execution Layer : Analyzing why "perfect" simulation cod...
博文
🚀 Week 4: Training the Brain – ML Integration & Poster Design Project: Autonomous Driving Supervisor: Roberto Ferrero Team: Yu Jiang, Runlai Yang, Haikun Xu, Jerry Fan 🧠 Machine Learning & Data-Driven Navigation This week, we successfully transitioned from classical control to Machine Learning (ML) implementation . We conducted controlled test runs on a predefined track to collect a comprehensive dataset of sensor readings and motion outputs. After cleaning the data to remove noise, we trained our path optimization model. The preliminary results are exciting: the model is now capable of supporting adaptive navigation, providing a much more "intelligent" response to the environment than simple logic-based code. 🎨 Professional Poster Preparation In parallel with our technical work, we began designing our Project Poster for the upcoming Week 5 submission. Our goal is to communicate our complex system architecture—from hardware integration to ML results—in a vis...
🚀 Week 3: Bridging Code and Movement – Integration Challenges Date: 15 February2026 Project: Autonomous Driving Supervisor: Roberto Ferrero Team: Yu Jiang, Runlai Yang, Haikun Xu, Jerry Fan 💻 Firmware Integration & Pathfinding Logic This week marked our transition from physical assembly to System Integration . We focused on developing the pathfinding algorithm and embedding it into the STM32 firmware. While we successfully brought several sensors under software control, integrating the complex pathfinding logic proved challenging. We encountered stability issues and unexpected behavior during execution, likely due to low-level hardware configurations. We are currently reviewing our implementation strategy to improve modularity and debugging efficiency. 🔧 Mechanical Reinforcement & Stability Motion testing revealed some "growing pains" for our Smart-Rover. As the car began to move, vibrations caused certain mechanical components to loosen. To ensure reliabil...
🚀 Week 2: From Parts to Platform – Full Assembly Complete Date: 6 February 2026 Project: Autonomous Driving Supervisor: Roberto Ferrero Team: Yu Jiang, Runlai Yang, Haikun Xu, Jerry Fan 🛠️ Full Mechanical Assembly & System Planning Great news—all required components have officially arrived! We kicked off the week with a final inventory check, confirming that our hardware "puzzle" is now complete. The highlight of this week was the Total Mechanical Assembly : we successfully mounted the motors, wheels, and specialized brackets for our sensors and control hardware onto the chassis. With the physical frame ready, we moved into the System Layout Planning phase. This structural foundation is essential for ensuring that our upcoming power distribution and signal wiring are both reliable and safe for operation. 💡 Weekly Reflection Completing the mechanical build taught us that a well-organized physical layout is half the battle. While we haven't encountered major ...
🚀 Week 1: Autonomous Driving Project Date: 30 January 2026 Project: Autonomous Driving Supervisor: Roberto Ferrero Team: Yu Jiang, Runlai Yang, Haikun Xu, Jerry Fan 🛠️ Hardware Inspection & Measurement During the first week, our primary focus was a thorough Hardware Inspection . We performed a detailed inventory check and conducted initial measurements of the electronic components—such as motor resistances and sensor voltages—to ensure everything met our design specifications. During this process, we discovered that several critical integration components were delayed in shipping, which prevented us from starting the full microcontroller-based setup. To stay productive, the team pivoted to the mechanical phase: we organized the available parts and completed the assembly of the vehicle’s chassis, including the DC motors and drivetrain. This ensured that our physical platform was stable and ready for future electronics. 💡 Weekly Reflection This...