Localisation machine learning #30
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The idea is to get a banch of VLX and robot position/orientation data using the simulator, and then establish a model to extrapolate the position of the robot using theses data (using linear regression, stochastic gradient descent...). The goal is to apply this model in the localisation_node to get and extra confidence of the real position of the robot (especially near walls)
Maybe get the dimensions of the robot and compute them in the future ?
Still not good enough for me, thinking of just using vlx near walls considering the orientation given by the odometry to choose which vlx to use in order to extrapolate position
PhileasL
commented
Feb 23, 2021
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Goal
The goal was to try to approach the position of the robot considering only the vlx values.
Approach
First of all, i used a supervised machine learning method instead of an unsupervised one because getting data in our case in very simple (and also because it's so much easier too). I used the NuSVR algorithm from sklearn python api.
Main problem
Here the problem was the multiality behavior of the solutions. For a given batch of vlx distances, multiple solutions are possible.
Solution
Splitting the game area in several distinct area. It has decreased the error but not enough for me.
Final performances
r² score of 0.9993, in a test of 6500 different positions and orientations, an average error of 6mm but a maximum error of 20cm.
And so what ?
We are not going to use theses models into our robots, it's not accurate enough for precise movement near walls in my opinion. The idea now is to create a small plugin which drive robots near walls considering only some vlx measures.