# sp_admiral

## Overview

The `admiral` node is responsible for the long term-planification.

To do so, the node updates an internal **score map** : a grid map in which each cell tracks the **interest score** of visiting the location: the more time since a drone last visited the area, the bigger the score to reap.

At a fixed rate, the node performs an optimisation pass to determine optimal targets for all active drones. It uses a custom **simulated annealing** ([definition here](https://en.wikipedia.org/wiki/Simulated_annealing)) process over the global system state `[x_a, y_a, x_b, y_b, ...]`.

## Outputs

![Admiral Orders](img/rviz_admiral1.gif)

You can see on the image above

- the occupancy grid of the area
- the score map:
  - blue : low score, cell was recently observed
  - red : high score, cell is more interesting to visit
- red cones : current sight areas
- green cones: target sight areas
- [A, B, C, D]: active drone ids