Obstacle Avoidance using Dynamic Movement Primitives

This project explores the concept of Dynamic Movement Primitives (DMPs), its extensions to encode Orientations and to avoid single and multiple point-like obstacles.

This project is an mixed implementation of the following papers:

1. Dynamical Movement Primitives: Learning Attractor Models for Motor Behaviors

2. Orientation in Cartesian space dynamic movement primitives

3. Movement reproduction and obstacle avoidance with dynamic movement primitives and potential fields

Results

Detailed information about the project can be found in here in the report

The Pybullet implementation of the work can be found below:

Sinlge Obstacle

https://user-images.githubusercontent.com/43097003/137211703-d2ad9737-4bb9-44b3-a769-d7329156d967.mp4

Multiple Obstacle

https://user-images.githubusercontent.com/43097003/137211729-6123f83f-cc04-408e-8d1b-d18a2b28343e.mp4