This PR handles different problems that were faced in the past such as the

• determining the orientation the robot needs to take, while applying the omnidirectional motion. NavFn planners and other Smac planners doesn't consider the omnidirectional constraints while planning a global path. Thus neo_mpc_planner, uses the points from the planner to generate a local path that would be feasible for omnidirectional motion.
• handling narrow path and avoids turns at the edges with additional lookahead distance and Distance Costmap generated based on bilinear interpolation.

Requirements:

• ESDF and the other related files for bilinear interpolation needs to be added to the python wrappers for nav2 costmaps.

Work to be done:

• Still there are issues while slowing down close to the narrow path. The footprint goes into collision with obstacles. This needs to be handled before the PR can be accepted. A PoC from simulation is ready. Needs to be verified with the real world scenario.