Case Study

Fixing navigation errors when a warehouse robot lacks boundary information

Case context: You are building an autonomous warehouse robot. It moves poorly in narrow aisles because the route-planning step needs aisle-edge locations, but the current pipeline does not provide them.

Question: According to good ML system design, what change should you make to the pipeline, and why is it better than forcing an existing module to do this job too?

Sample answer: Add a dedicated module that detects aisle edges and sends that information to the route-planning step. This is better because it supplies the missing signal directly and keeps the other modules simpler to design and train.

Key points:

  • Add a dedicated aisle-edge detector.
  • Pass the needed boundary information to route planning.
  • Avoid overloading existing modules with extra complexity.

Rubric: The response should identify adding a new boundary-detection module and explain that it gives the planner the missing information while avoiding unnecessary complexity in other parts of the pipeline.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI

Related