Kris' changes

[abhranildas]

• Classify with ray_scan region, i.e. add samp_correct to example.

• Add the line example with ray_scan, quad (gx2) and fun alternatives. See how different the results are.

• Perhaps add something to the performance section about the above comparison: Re-define relative diff. in terms of ray-scan result. Make two columns, rel. diff and time, and sub-columns for each method, including fun. NOPE

• Add fun to fig. 4a. NOPE

• Maybe in ms text, add: for example, the ray_scan function of a line at y=k returns...

• Compare two examples in demo: combining the circle regions using a function, vs ray-scanning quads, then using the combination operators. The second ought to be more accurate.

• Then add back the text for these domain combination operations in ms, to say why they are useful.

• Add set operation example figure to the bottom of 2d.

• Add this in the performance section to compare between ray-scan and fun. NOPE

• Documentation for custom ray_scan function etc. init_sign is at -infty.

• Go through his e-mail again.

• C++/MEX implementation

• Add topics to intro: robotics, chance constrained programming/planning, probabilistic safety, risk-aware planning, add his citations.

• Mention to Kris that I acknowledged him