Some Frequently asked questions related to low paths in the path
planning section of the robotica practical.


> What is low path good for in practice?

The purpose of the low path exercise is to teach you how to do path
planning in a discrete grid. So it is primarily a teaching
exercise. Sometimes robots work in a workspaces with obstacles. If the
obstacles are large path planning is necessary. In our case it is easy
to avoid the obstacles by doing high paths. We ask you to do the low
paths anyway as an exercise. 

Furthermore, the low paths are executed faster by the robot than high
paths since the robot raises the arm less and therefore travels a
smaller total distance. This can also be seen as a practical
justification.

> In the case where there is an obstacle, do we use first 
> lowpath, then highpath and then low again? 

It is a design choice. I leave it up to you. But to make things simple
I suggest that if a low path is not possible due to an insurmountable
obstacle to generate a high path from the start to the goal position.

> Is lowpath just for the pown and the tower and sometimes for the queen 
> and for the king?

Low paths should be performed for all pieces for which a low path from
the start to the goal position can be found using 4-connectivity.

4-connectivity
    2
 1  p  3
    4

contrast this to 8-connectivity
  1  2  3
  4  p  5
  6  7  8


So no diagonal low-paths please (e.g bishops). They will not work.
Forget about 8 connectivity and move bishops,queens,pawns etc using 
paths with 4 connectivity.


> Do we have to implement a search algoritm?

It depends what you understand by search. I have implemented it with a
distance transform (see Leo's notes). Some think that this is not a
real search since it does not explore a tree until a solution is
found. But there are other ways to implement it which look like
tree-search. It is up to you to decide how you want it programmed.

So long as it works it is fine.