Open Zach10a opened 3 years ago
A naive way of approaching this would be to simply generate the risk map and find a path for every timestep, which would be computationally infeasible currently. Some form of time constraints specified by the user could allow the number of runs to be reduced significantly. Further use of some form of heuristic wold likely be valid as well (eg. daytime is always assumed to be busier than night, therefore only generate for night time if minimising risk)
Implement a system whereby the user is able to see how the ground risk of a UAV changes over time. This could take the form of a min/max risk output, when they occur and how the risk changes over time.