Transitioned to a gradient-descent only solution

Added empirically derived fudge factor to fitness function
Eliminated dangerous spurious local minimum near the origin.
Now taking into account the normals
No longer using any of the point cloud stuff
Huge speed improvements
Huge memory usage improvements
Added code smell.  Will need to deodorize later.

1 files changed, 3 insertions, 0 deletions

diff --git a/tools/lighthousefind_tori/main.c b/tools/lighthousefind_tori/main.c
index aa51448..ee56b37 100644
--- a/tools/lighthousefind_tori/main.c
+++ b/tools/lighthousefind_tori/main.c
@@ -51,6 +51,9 @@ static void runTheNumbers()
 			to->sensor[sensorCount].point.x = hmd_points[i * 3 + 0];
 			to->sensor[sensorCount].point.y = hmd_points[i * 3 + 1];
 			to->sensor[sensorCount].point.z = hmd_points[i * 3 + 2];
+			to->sensor[sensorCount].normal.x = hmd_norms[i * 3 + 0];
+			to->sensor[sensorCount].normal.y = hmd_norms[i * 3 + 1];
+			to->sensor[sensorCount].normal.z = hmd_norms[i * 3 + 2];
 			to->sensor[sensorCount].theta = hmd_point_angles[i * 2 + 0] + LINMATHPI / 2;
 			to->sensor[sensorCount].phi = hmd_point_angles[i * 2 + 1] + LINMATHPI / 2;
 			sensorCount++;