Comprehensive Engineering Consultation and Research Strategy for High School Go-Kart Project

Mentor Background: Brian graduated with a bachelor’s degree in Aerospace Engineering from Purdue University, where he participated in the College of Engineering honors program. He is currently a graduate student specializing in Artificial Intelligence ( $AI$ ).
Student Profile: Aaron and Aurelia are students at an Asian high school who have built a go-kart. Aaron is interested in Aerospace and Mechanical Engineering, while Aurelia is considering Mechanical, Aerospace, or Electrical Engineering.
Meeting Goal: This is an informal, free initial consultation with the primary intent to discuss engineering research values, refine project ideas for the Orange County Science and Engineering Fair ( $OCSEF$ ), and establish a structure for an engineering project.

Public vs. Industry Perspective: Brian highlights a discrepancy between what the public thinks engineers do and the actual daily tasks. * Public Perception: Often revolves around the merit of a device (e.g., "rockets are cool") or passion-based builds. * Industry Reality: Engineering in industry focuses on meeting specific needs of clients, customers, or end-users, even if those users cannot articulate their needs.
The Henry Ford Case Study: When Henry Ford invented the car, he noted that if he had asked people what they wanted, they would have said "a faster horse." Engineers must extrapolate the true objective (’getting from point $A$ to point $B$ faster’) rather than accepting the customer’s proposed solution.
Under-defined Problems: Problems in engineering are rarely like math class with a single solution. They are often under-defined, requiring engineers to collect substantial information to determine the direction of a solution and justify that direction with data.

The Bridge Metaphor: Research is described as a "continuously evolving discussion." Beginners often feel they must create something entirely new from scratch ( $0$ to $1$ ). In reality, most researchers walk "three quarters out on the bridge that is already built" by others and then add their own small contribution.
Literature Reviews: Identifying gaps in existing research is essential. Students are encouraged to read the "conclusion" and "discussion" sections of peer-reviewed papers, as authors frequently list lingering questions or areas for future investigation.
Academic Impact Example: Brian mentions his undergraduate project involving $AI$ for guided parachute delivery systems. The project was motivated by the fact that the $US$ lost approximately $33,000,000$ in humanitarian and military aid during the Bosnian War because systems were failing. He added a single technique to existing research, providing a significant evaluation of real-time sensor data, which led to a publication and helped secure his current job.

1. Broad Goal: Identify the general aim (e.g., getting from point $A$ to point $B$ faster).
2. Customer Needs (Qualitative): Identify non-technical requirements such as safety, comfort, usability, reliability, controllability, and user-friendliness.
3. Technical Requirements (Quantitative): Translate qualitative needs into measurable units. * Example: If the need is "Safety," a technical requirement could be the number of $G$ -forces ( $g$ ) a human can survive in a crash.
4. Target Value: Set a realistic goal for the system to meet. * Example: Ensuring the passenger experiences no more than $8.00\,g$ during a collision.
5. Cost-Benefit Analysis/Optimization: Justify over-designing a system. If reducing forces to $4.00\,g$ doubles the car’s price but only increases survival rates by $0.01\%$ , the trade-off may not be valuable to the customer.

Research Databases: * ProQuest: A comprehensive academic database Brian used from middle school through college. * arxiv.org: An open-access database hosted by Cornell University for research papers.
Artificial Intelligence Platforms: * LLMs for Research: Brian recommends using Claude (Anthropic) for brainstorming and summarizing papers over Gemini (Google) or ChatGPT (OpenAI), though he warns that claims must be validated manually $40\%$ of the time. * PyTorch: The industry-standard open-source platform for building neural networks and $AI$ models. Brian used this to build predictive models for the HEM 190 Black Arrow offensive missile system.
Programming and Hardware: * Arduino: Mentioned for hardware control. * Python: A foundational language for $AI$ and machine learning.

The Testbed Approach: Use the go-kart not just as a vehicle, but as a "physical hardware testbed" to validate software-based computer models. This is highly valued because many researchers only work in simulation.
AI-Enabled Predictive Traction Management: * Inspired by the Moose Test (lateral grip assessment). * The system could assess a driver's performance and provide real-time warnings or analysis regarding understeer, oversteer, and traction limits. * This targets an "Amateur Racer" or "Off-roader" client base who needs to handle specialized terrain safely.
Computer Vision Coaching: References a $2021$ $OCSEF$ project that used computer vision as an archery coach as a precedent for personal training tools.
Efficiency Metrics: Testing acceleration vs. voltage drop ( $V$ ) or predictive accuracy vs. runtime. Brian shares an example of improving a model's runtime by $3$ orders of magnitude ( $1,000 \times$ faster) while losing less than $2\%$ accuracy.

Pre-approval Notice: Projects involving "Potential Hazards" (motorized equipment, high-powered mechanical systems) or "Human Subjects" (drivers) must submit safety forms for approval before experimentation begins.
Deadlines: * The pre-approval deadline mentioned is November 3rd. * Final project submission is typically in February.
Safety Compliance: Testing must follow state and local laws. This may require testing on private property (e.g., a desert location or a paved parking lot with permission) rather than public streets.
Project Planning: Engineers use Gantt Charts to visualize the timeline for deliverables, ensuring the project aligns with school and fair requirements.
Student Commitment: The students plan to dedicate approximately $45.00$ to $60.00\,\text{minutes}$ per day over the summer to the project.

Weekly/Daily Schedule: Brian will help students implement a project plan for the fall semester.
Drafting Engineering Specs: Students will refine their hardware/software goals using the five-step process.
Resource Sharing: Brian will provide links to arxiv.org, annotated bibliographies, and specific research directions based on the go-kart’s off-road nature.