Jump Point Search was not actually devised for orthogonal search on uniform cost grids. It actually takes advantage of heading diagonally (over straight directions) to reach a goal.
In the actual implementation of JPS, the problem was partially solved, but not in a very elegant manner, as it floods the search with forced neighbors, jump points and discards evrey single diagonal expansion. It works fine, but the search time is a bit long (higher than standard A-star) and the return path is a step-by-step path.
The point here is to look into ways to develop a new set of pruning rules for 4-directions grids. Some starting points that might be worth investigating further or detailed articles that can be used for reference:
Jump Point Search was not actually devised for orthogonal search on uniform cost grids. It actually takes advantage of heading diagonally (over straight directions) to reach a goal.
In the actual implementation of JPS, the problem was partially solved, but not in a very elegant manner, as it floods the search with forced neighbors, jump points and discards evrey single diagonal expansion. It works fine, but the search time is a bit long (higher than standard A-star) and the return path is a step-by-step path.
The point here is to look into ways to develop a new set of pruning rules for 4-directions grids. Some starting points that might be worth investigating further or detailed articles that can be used for reference: