Safe Online Data Driven Learning

Posted on August 12, 2015

For more than a year now I have been working with my collaborator Paul Lutkus on control barrier function generation for an autonomous robot agent moving in an unknown 2-dimensional environment using discrete data driven optimization. This project is soon to be published in conference_name.

This post is still a work in progress and will be updated as and when new results insight is obtained and ideas are discussed. To read about Paul’s take on the same, be sure to check out his blog.

Table of Contents

  1. Problem Statement
  2. Control Barrier Functions
  3. Data driven learning
  4. A toy example
  5. The dual objective
  6. Modeling obstacles - the Alpha shape References

Problem Statement

Instead of writing this out asbtract style, here is a list of things that we had to do and most of which we have implemented as part of the project:

References