Updating SVM Stuff and removing sponsors

documentation/html/source/index.rst

@@ -118,15 +118,6 @@ Loading Data
    terminal and press a key in the ``cv2`` window or continue pressing
    keys until all images are cycled through.
 
-
-Sponsors
-========
-
-None of this would be possible without our absolutely amazing sponsors!
-
-.. image :: ../../index_images/sponsors.png
-    :align: center
-
 .. toctree::
    :maxdepth: 2
    :caption: Contents:

documentation/html/source/perc22a/svm/SVM.rst

@@ -10,12 +10,21 @@ SVMs are a discriminative classifier: that is, they draw a boundary between clus
 midline that best divides the blue and yellow cones from the Perceptions pipeline. This midline is the SVM decision boundary.
 We use this midline as the reference line for the lane center, and the primary input to the Control pipeline.
 
+The SVM algorithm is part of the the path planning pipeline, which is responsible for controlling the car. The SVM algorithm, however, is
+written in Python and for the purposes of documentation, is included in the Perceptions documentation.
+
 How does it work?
 -----------------
 
 The SVM algorithm takes in the cones class, which contains the x and y coordinates of 2 types of cones, the blue and yellow cones (it also takes in
 orange cones, but for the purposes of the current SVM, they are not a primary concern). The SVM algorithm then fits a line to the data, which is the
 SVM decision boundary. This line is then used as the reference line for the lane center. 
 
-@DAN - I think this is a good place to start. I think we should start by explaining the SVM algorithm, and then move on to the SVM implementation in the files
-Also maybe pictures or diagrams? Idk just a thought
+Our Perceptions pipeline outputs the x and y coordinates of the blue and yellow cones. These x and y coordinates are given in global frame
+every 0.1 seconds. The SVM algorithm takes in these x and y coordinates and fits a line to the data. It does this by taking the mesh frame from the
+cones and passing a line through the center of the mesh frame. This line is the SVM decision boundary. After we get a rough estimate for 
+what the decision boundary is between the 2 colors of cones, we then downsample the points on the line to provide an equidistant line through
+all of the cones in the mesh frame as a better center line for the car.
+
+Additionally, we also provide 2 base cones behind the cars original location to provide the SVM with an understanding of color and distance.
+After we get the decision boundaries from the SVM, we then use the decision boundaries to improve coloring for our LiDAR algorithm.