Engine-RPM-Profile-Detection

Highlights of the project

• Figure out real world problems

• Algorithm was created by ourselves

• Satisfied output

• Collaborative environment

1. Introduction

Sponsor

Main Purpose: To find 2 representative indicators, HighRPM and MaxRPM, with certain statistical significance.

2. Data Inspection

3. Approaches

• 3.1 Summarization Approach

  • Analyze each single trip of a specific platform.

  • Summarize the results from all trips of a specific platform to predict RPM profile of the platform.

• 3.2 Regression Approach

  • Find the inflection point of the torque from maximum value.

3.2.1 Main steps

• Step 1: Find the contour curve of data

• Step 2: Introduce linear regression

  • Fitting N lines of linear regression

  • Horizontal line of the average value of top K of torque

  • N intersections among these two types of lines

• Step 3: Training the model and determine the optimized intersection.

  • Ground truth: The theoretical value of HighPRM in handbook

  • Supervised learning to determine N and K

  • The intersection of two lines will be the predicted value.

• Step 4: Statistical significance

4. Development and Collaboration Environment

5. Code Structure

6. Paperworks

Progress report