Case Study

Resolving Path Planning Failures in a Self-Driving Vehicle

Case context: You are developing a self-driving pipeline, and your vehicle struggles to position itself properly on roads without clear boundaries. You realize that a skilled human driver needs to know the location of the lane markings, but your current pipeline lacks a module to output this information to path planning.

Question: Based on Andrew Ng's guidelines, what architectural modification should you make to the pipeline to address this issue, and why is this choice superior to embedding this task into existing modules?

Sample answer: You should modify the pipeline by adding a dedicated "Detect lane markings" component. This is superior because it directly supplies the missing lane-marking information to the path planning module while preventing other modules from becoming overly complex to build and train.

Key points:

  • Add a dedicated "Detect lane markings" component to the pipeline.
  • Provide the missing lane-marking location information to the path planning module.
  • Prevent existing modules from becoming overly complex to build or train.

Rubric: The response must identify adding a lane-marking detector module to the pipeline. It must explain that this resolves the missing information issue for path planning and avoids making other pipeline modules overly complex to build or train.

0

1

Updated 2026-05-26

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Related