Simulate rover movement on a grid

You are building a rover navigation simulator.

Part 1: Single rover on a 2D map

You are given:

• A rectangular 2D grid with coordinates (x, y) .
• A rover starting position (x0, y0) and a facing direction (one of N, E, S, W ).
• A command string consisting of:
  • L : turn left 90°
  • R : turn right 90°
  • M : move forward 1 step in the current facing direction

Rules:

• The rover cannot move outside the map boundaries. If a command would move it out of bounds, ignore that move (or alternatively, return an error—state your chosen behavior clearly and implement it consistently).

Task: Implement a function that returns the rover’s final position and direction after executing all commands.

Part 2: Multiple rovers

Extend the design to support multiple rovers on the same 2D map.

• Each rover has its own initial state and command string.
• (Clarify your assumption) Either rovers execute sequentially (one after another) or simultaneously step-by-step.
• Decide and implement a collision policy (e.g., rovers cannot occupy the same cell; a move that would collide is rejected).

Task: Explain how you would change the code structure/API to support multiple rovers cleanly and safely.

Part 3: 3D map

Now the map is 3D with coordinates (x, y, z).

• Define a reasonable model for orientation in 3D (e.g., a heading direction plus pitch, or facing as one of 6 axes: +X, -X, +Y, -Y, +Z, -Z ).
• Extend commands accordingly (you may introduce new commands if needed, but specify them).

Task: Describe and/or implement how your solution generalizes to 3D while keeping the design maintainable.