Write tutorial on using Beluga

[hidmic]

Feature description

Precisely what the title says. While Beluga's API is well documented, the only reference on how to make use of the library is a ROS node and some tests. That works against adoption. We need a golden example that builds up in complexity, showcasing Beluga's features. We need to guide newcomers on reproducing that example, one step at a time.

[nahueespinosa]

Following up on the conversation on how to implement the golden example, the idea is to implement the one-dimensional hallway example for Monte Carlo Localization presented in Chapter 8.3 in Thrun et. al. using the Beluga core library without ROS dependencies.

Here is a list of milestones that can be used as reference:

1. Implement the program skeleton, focusing on the creation of a particle set (states and weights) structure and basic visualization capabilities.
   • This can be serializing to a file and then post-processing in Python, or simply writing the output to the terminal in a way that resembles the example in the book.
   • It is key to represent both the position (state) of the particle and the weight (could be a line's height, a circle's diameter, or something else).
2. Develop the initialization function for the particle set to sample uniformly from a specified interval.
   • See std::uniform_real_distribution and beluga::views::sample.
3. Implement the sensor model capable of reweighting particles based on the map landmarks (doors) and the current position.
   • This can use a simplified version of the landmark model proposed in Chapter 6.6.
     $w = e ^ {{- d^2} / {2 \sigma}}$ where d: distance to the door
   • See beluga::sensor and beluga::actions::reweight
4. Develop the motion model to update the positions of the particles, including a routine for moving particles at a constant rate in a given direction.
   • See beluga::motion and beluga::actions::propagate.
5. Implement a basic resampling step that allows for the selection of particles based on their weights (no need to make it adaptive at this stage, so taking samples up to a fixed size is perfectly fine).
   • See beluga::views::sample again.

FYI @Alondruck @hidmic @gonzodepedro

[hidmic]

Addressed as of #365 and #387. @Alondruck are the scripts on #387 used for our tutorial?