# \[Nav2 WG\] August 31, 2023

**URL:** <https://discourse.openrobotics.org/t/nav2-wg-august-31-2023/33195>\
**Category:** Navigation\
**Created:** [August 31, 2023, 7:57pm UTC](https://discourse.openrobotics.org/t/nav2-wg-august-31-2023/33195 "2023-08-31T19:57:53Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![smac](https://sea2.discourse-cdn.com/flex022/user_avatar/discourse.openrobotics.org/smac/32/6763_2.png) [@smac](https://discourse.openrobotics.org/u/smac)\
**Post date:** [August 31, 2023, 7:57pm UTC](https://discourse.openrobotics.org/t/nav2-wg-august-31-2023/33195/1 "2023-08-31T19:57:53Z")

</div>

Navigation Working Group Meeting Minutes - August 31

Attendees:

- Steve Macenski
- Alexey Merzlyakov
- Ryan Friedman
- Awab Syed
- Reza Kermani
- Magda Skoczen
- Josh Wallace

New people introduce themselves

- Magda - join meeting + some questions
- Reza - construction AMR company engineer

Topics

Announcements:

- N/A

New additions of note:

- Wait recovery now respects ROS time (Dexory)
- New changes to Smac Planner complete to add new features for (1) additional search primitives (2) optional downsampling of cost heuristic representation and (3) quadratic cost functions. Testing would be appreciated in PR #3752
- Humble fix for planner server lifecycle management

Ongoing:

- GPS tutorial coming shortly (Kiwibot) [GPS WPF tutorial by pepisg · Pull Request #457 · ros-navigation/docs.nav2.org · GitHub](https://github.com/ros-planning/navigation.ros.org/pull/457)
- Interpolating lookahead point at the goal pose to get to the goal pose after in RPP (James Ward)
  - Alexey: Prune distance lookahead parameter for waypoint follower related?

New bugs of note:

- Bug identified by Alexey [rclcpp\_action goal handle timeout logic clearing long-running goals · Issue #3765 · ros-navigation/navigation2 · GitHub](https://github.com/ros-planning/navigation2/issues/3765#issuecomment-1689951628)
  - rclcpp actions results not properly returned due to timeouts in rclcpp’s handling of goal handles
  - Alexey: comparing time with timeout, the goal is killed. Check when last updated, not when the goal is initiated.
  - Steve: Oh I like that better than checking completion, in case there’s an issue in completion logic of user code. Lets open a ticket in rclcpp and ping me and Chris.

Anything else?

- Alexey link to ROS Developer’s Day talk:

[![](https://us1.discourse-cdn.com/flex022/uploads/ros/original/3X/f/8/f83a49b49a99821d616ada2f03c8223f6eb6c710.jpeg "Collision Monitor – Software-Based Safety Module | Alexey Merzlyakov | ROS Developers Day 2023") ](https://www.youtube.com/watch?v=bWliK0PC5Ms)

Round Robbin (questions, comments, what youre working on, updates, etc)

- Steve: Smac Planner remove zig-zags caused in Hybrid-A\* due to angle quantization bin skipping (+ a bunch of new features!), creating ROSCon talk on MPPI controller
- Alexey: Working on the WPF flaky tests (1) rclcpp action (2) controller prune issue (3) BT errors not being cleared out. Working on rclcpp::Rate
- Awab: Nav and having problems without AMCL → answered questions.
- Josh: Route server having problems with routing potentially, working on finding closest node
- Magda: Interested in costmap pub/sub updates with questions, awesome!
- Ryan: Got Nav2 working with Cartographer on a quadcopter, costmap issues took some time to work through. Working to move to Humble

---

<div class="post-metadata">

**Author:** ![smac](https://sea2.discourse-cdn.com/flex022/user_avatar/discourse.openrobotics.org/smac/32/6763_2.png) [@smac](https://discourse.openrobotics.org/u/smac)\
**Post date:** [September 4, 2023, 6:22pm UTC](https://discourse.openrobotics.org/t/nav2-wg-august-31-2023/33195/2 "2023-09-04T18:22:35Z")

</div>

One thing from the title of Alexey’s talk icon I feel compelled to clarify is that the Nav2 Collision Monitor is **not** a certified safety system with any guarantee or remark about meeting any safety standards. It uses the same kinds of techniques as some certified systems, but it is a pure-software and uncertified solution and should not be interpreted as providing formal safety promises.

Just to make sure that this is abundantly clear 🙂 Its absolutely useful and a great addition to the stack for those without safety needs or without certifiable safety sensors to provide an additional monitor to prevent collision. There’s also been some work recently adding dynamic footprints for changing vehicles and a collision detector node which rather than impacting the robot’s state, publishes to a topic for use by another system to perform an action (e.g. trigger LEDs or ignore areas where the work-product is within the monitor zone, etc)