Jazzy Nav2 SLAM - map generation on the Pi5;-)

[Marco Walther]

It looks like the Pi5 is fast enough to run Nav2, at least the map
generation half of it;-)

The slam toolbox seems to have some problem with the /scan (?) messages,
It can't keep up. That's still something to figure out?! Some of the
10Hz messages seem to make it through.

[sync_slam_toolbox_node-12] [INFO] [1763779023.501778754]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779023.401 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779024.101634552]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779024.001 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779024.701861026]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779024.601 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779025.301596970]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779025.201 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779025.902261784]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779025.801 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779026.501532244]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779026.401 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779027.101571927]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779027.001 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779027.702235862]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779027.601 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779028.301529263]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779028.201 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779028.901680607]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779028.801 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779029.501657014]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779029.401 for reason 'discarding message because the queue
is full'
[sync_slam_toolbox_node-12] [INFO] [1763779030.101971604]
[slam_toolbox]: Message Filter dropping message: frame 'base_laser_link'
at time 1763779030.001 for reason 'discarding message because the queue
is full'

Have fun,
-- Marco

[Sergei Grichine]

Marco,

Two of my robots (Seggy and Dragger) run on RPi5 8Gb, Turtle runs on RPi4 8GB overclocked to 2 GHz.

Neither have a problem running whole stack (SLAM Toolbox and Nav2). Turtle is at ~65% CPU though.

Maybe, loop frequencies are too high? My parameters and everything is at  https://github.com/slgrobotics/articubot_one/tree/dev

Wiki at https://github.com/slgrobotics/articubot_one/wiki  with more detail.

I'll look into this more soon, maybe I just ignore some warnings as all seems working just as it should.

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/968bf18f-f5de-4114-be06-74bd177bd9de%40gmail.com.

--

Best Regards,

-- Sergei

[Pito Salas]

Marco

I have been digging into a very similar scenario but on a Raspi4. I have some questions:

1) Are you also seeing: "[INFO] [planner_server-3] [2025-11-22 10:19:20.748407]
Message Filter dropping message: frame 'laser' at time 1763824760.440 for reason 'the timestamp on the message is earlier than all the data in the transform cache’”

2) Are you launching slam_toolbox with it’s “default” launch and yaml files and if not what have you tweaked?

3) Do you have rviz running at the time because I’ve seen that have an impact

4) Why are you running sync_slam_toolbox? I am running async. Based on what I’ve read I thought that one would be more tolerant of slowness in the pipeline.


Best,

Pito

Boston Robot Hackers &&
Comp. Sci Faculty, Brandeis University (Emeritus)

> --
> You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/968bf18f-f5de-4114-be06-74bd177bd9de%40gmail.com.

> <Screenshot_20251121_184402.png>

[Sergei Grichine]

Marco,

I was running Seggy around today, "cold" SLAM Toolbox mapping (no known map initially). Whole stack onboard, as usual, only RViz on my Desktop.

The RPi 5 "htop" shows no overload:

The mapping process, using Nav2 goals, with very little help from the joystick, went smoothly.

Seggy even went through the narrow (79 cm) hallway all by itself (due to my recent tuning of footprint etc.).

I attached a full console log, in case you might find some resemblance to what you are observing.

My code and Wiki is, of course, here: https://github.com/slgrobotics/articubot_one/wiki  ("dev" branch most current) 

Best Regards,

-- Sergei

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/541AEFCE-4C72-43E9-ACFF-918D7CEC8255%40gmail.com.

[Pito Salas]

Dear Sergei


Thanks for the valuable use case and examples!

Quick questions:

Where is your slam_toolbox.yaml? Or are you using Cartographer?

Also why the delayed_include?

Best,

Pito

Boston Robot Hackers &&
Comp. Sci Faculty, Brandeis University (Emeritus)

> On Nov 22, 2025, at 12:28 PM, Sergei Grichine <vital...@gmail.com> wrote:
>
> Marco,
>
> I was running Seggy around today, "cold" SLAM Toolbox mapping (no known map initially). Whole stack onboard, as usual, only RViz on my Desktop.
>
> The RPi 5 "htop" shows no overload:

> <Screenshot from 2025-11-22 11-01-22.png>
> The mapping process, using Nav2 goals, with very little help from the joystick, went smoothly.
> Seggy even went through the narrow (79 cm) hallway all by itself (due to my recent tuning of footprint etc.).
>

> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/CA%2BKVXVPoAndZMVd_4J4T5kS2eSQ7pz_3Dv1ASvndjoEcHw6FLA%40mail.gmail.com.
> <ros_seggy_launch_bootup.zip>