[Bug]: Trajectory provided sometimes starts at origin when leaving lift

[aaronchongth]

Before proceeding, is there an existing issue or discussion for this?

• I have done a search for similar issues and discussions.

OS and version

Ubuntu 22.04

Open-RMF installation type

Source build

Other Open-RMF installation methods

No response

Open-RMF version or commit hash

main

ROS distribution

Humble

ROS installation type

Binaries

Other ROS installation methods

No response

Package or library, if applicable

No response

Description of the bug

While running the hotel world simulation demo, by running robots up and down the lifts, there is the rare chance that a trajectory segment starting from (0, 0) or (0, #) or (#, 0) will be produced. It is hard to reproduce and goes away on it's own in a few seconds.

Sometimes only either x or y values will be 0, while some times both are 0.

Steps to reproduce the bug

Hard to reproduce

1. Run simulation demo of hotel world (perhaps add names to waypoints near the lift to make repeating going up and down faster)
2. Run a patrol task between a waypoint on L1 and L2 with a great number of loops
3. When a robot trajectory is generated to respond to a request, where the trajectory starts from the lift waypoint, it has a small chance of being produced.

I logged it a little using https://github.com/open-rmf/rmf_visualization/tree/ac/zero-segment

Expected behavior

Trajectory segments should start from the lift waypoint coordinates and not anywhere with a zero coordinate.

Actual behavior

Sometimes the first segment of the trajectory starts from (0,0), which causes a visualization of the rogue trajectory

Additional information or screenshots

Sorry for the bad ss quality,

How we get the trajectory from the dashboard, this is the case where only x was 0

weird_traj-2023-11-02_14.02.10.mp4