An active area in research and industry is managing robotic swarms for precision agriculture. This research focuses on the particular task of path planning for agents in a swarm. Given a known map of an agriculture field, the goal is to allocate and create paths for the agents so that they can visit all the targeted crops in an efficient manner as a team. This problem is NP-hard and is closely related to the vehicle routing problem (VRP) in operations research.
The following research questions were asked:
1. What are the most efficient algorithms for generating paths for a fixed number of particle agents in a 2D environment with known static interest points?
2. How effectively can paths be planned through obstacles representing typical agriculture landscapes?
3. What are the optimal parameters for minimising time and distance travelled for
these agents?
4. How does the distribution of the targeted points influence the paths? Specifically,
if the task is to visit all points, randomly scattered points, or points scattered
in clusters?
To answer these questions, two types of field models (continuous and grid) were created in Python and several different path planning algorithms were tested. The results show that the swarm size is very influential, and also that there is a compromise between optimising for metrics and the quality of the solution.
