Robot Navigate

Trigger: The user asks the robot to map a new environment, or the existing map needs to be rebuilt.

Workflow:

1. Capture the current camera frame.

2. Inform the user: "Mapping is about to begin. Are you ready?"

3. Wait for the user to confirm they are ready.

4. Start mapping:

   uv run {baseDir}/scripts/build_map.py start

5. Inform the user: "Mapping has started. Please use the remote control to drive the robot in a complete loop around the area to be mapped. Let me know when you are done."

6. Wait for the user to confirm the robot has completed the loop.

7. End mapping:

   uv run {baseDir}/scripts/build_map.py end

8. Inform the user that mapping is complete and they can proceed to pose initialization.

---

## Sub-function 2: Pose Initialization (Relocation)

**Trigger:** After mapping is complete and before first use of navigation, or when the robot has lost its position and needs to relocalize.

**Workflow:**

1. Capture the current camera frame。

2. Start relocation:
```bash
   uv run {baseDir}/scripts/init_pose.py eth0

3. Read the returned result:
   • statusCode == 0 → initialized is saved as true in slam_state.json. Inform the user that initialization succeeded.
   • statusCode != 0 → Inform the user: "The robot is not within the mapped area. Please move the robot back into the mapped area and try again." Do not proceed with any navigation operations.

Sub-function 3: Mark Navigation Waypoint

Trigger: The user says "I want to add a new navigation waypoint" or similar.

Prerequisite: Pose initialization (sub-function 2) must have been completed. If not, prompt the user to run initialization first.

This sub-function involves a multi-turn dialogue, as follows:

1. Capture the current camera frame。

2. Robot replies:
   "Please make sure you are within the mapped area, then use the remote
    control to drive the robot to the position you want to mark. Let me
    know when you are done."

3. Wait for the user to confirm (e.g. "Done", "I've marked it", "OK", etc.).

4. Robot replies:
   "Please give this waypoint a name."

5. Wait for the user to provide a name (e.g. "fridge").

6. Record the waypoint:

   uv run {baseDir}/scripts/record_goal.py <goal_name>

7. Read the returned result:
   - Outputs "已保存目标点: <name>" → Inform the user the waypoint was saved,
     e.g. "The current position has been marked as 'fridge'."
   - Outputs "获取当前位置失败" → Inform the user that marking failed.

Sub-function 4: Navigate

Trigger: The user asks the robot to go to a previously marked waypoint.

Workflow:

1. Check prerequisites — get current SLAM state:

   cat {baseDir}/scripts/slam_state.json

Confirm has_map == true and initialized == true. If any condition is not met → Inform the user which prerequisite is missing.

2. Check target waypoint — list available goals:

   uv run {baseDir}/scripts/nav_to_goal.py --list

Confirm the target waypoint is in the returned list. If not → Inform the user the waypoint does not exist.

3. Capture the current camera frame。

4. Start navigation:

   uv run {baseDir}/scripts/nav_to_goal.py --target <goal_name>

5. Read the returned result:

   • Outputs "已到达" / is_arrived: true → Inform the user the robot has arrived at the target.
   • Navigation fails → Report failure and reason to the user.

6. Collision handling (built into nav_to_goal.py):

   • Pause navigation → Query obstacle service → If clear: advance and resume; if blocked: turn 90° and retry.
   • If all four directions are blocked: report failure to the user.

Sub-function 5: Get SLAM State

Trigger: Whenever the current navigation system state needs to be queried, or as a prerequisite check before executing navigation.

Workflow:

1. Capture the current camera frame。

2. Get SLAM state and goal list:

   cat {baseDir}/scripts/slam_state.json
   uv run {baseDir}/scripts/nav_to_goal.py --list

3. Report the current navigation system state to the user.

State fields in slam_state.json:

Sub-function 6: Delete Current Map

Trigger: The user asks to delete the current map, clear all navigation data, or start fresh.

Workflow:

1. Warn the user: "This will permanently delete the current map and all saved waypoints. Are you sure you want to continue?"

2. Wait for the user to confirm.

3. Delete the map state and waypoint files:

rm -f {baseDir}/scripts/slam_state.json
rm -f {baseDir}/scripts/poses.json

4. Read the result:
   • Both files deleted successfully → Inform the user: "The current map and all waypoints have been deleted. You will need to run Mapping and Pose Initialization before navigating again."
   • Deletion fails (e.g. permission error) → Report the error to the user and do not proceed.

Notes

• Do not rewrite any output prefix format.
• The waypoint marking flow must follow the full multi-turn dialogue. Do not skip user confirmation steps and call the underlying script directly.
• If relocation fails, the user must be prompted to manually move the robot back into the mapped area. Do not retry automatically.
• Navigation is autonomous — the robot will plan a path and move to the goal on its own, with automatic collision avoidance (pause, check obstacles, turn or advance, resume).
• Use record_goal.py to save the robot's current position as a new named goal, with the goal stored in {baseDir}/scripts/poses.json.
• Use --list to dynamically view all available goals from poses.json.