Nav2 for AMRs — Mastery Plan

For

Engineer who already understands basic ROS 2 nodes, TF2, and the high-level Nav2 flow, but wants deep operational ownership of navigation on a real AMR.

Goal

Understand Nav2 end to end: architecture, lifecycle, costmaps, planners, controllers, recovery logic, localization contracts, AMR-specific constraints, debugging, tuning, and extension points.

Why This Track Exists

The current ROS 2 track gives you the architectural overview, but AMR-grade Nav2 work requires a separate layer of depth:

• You need to reason about why a robot stops, oscillates, replans forever, or exhausts recoveries.
• You need to connect TF, localization, costmaps, planners, controllers, and BT execution as one system.
• You need to debug from logs, bags, topics, parameters, and field symptoms, not from theory alone.
• You need the judgment to tune and extend Nav2 for warehouse, factory, and aisle-constrained robots.

This track is designed for that jump from “I know what Nav2 is” to “I can own Nav2 in production.”

Prerequisites

• 01 — Nodes Topics Actions
• 02 — Tf2 Time Qos
• 03 — Nav2 Architecture

Before starting this track, you should already be comfortable with:

• ROS 2 actions, lifecycle nodes, and parameter files
• TF tree semantics: map -> odom -> base_link -> sensor_frame
• Basic costmap, planner, and controller vocabulary
• Reading ROS 2 logs, bag files, and topic graphs

Planned File Structure

nav2/
├── 00-learning-plan.md
├── 01-nav2-system-architecture.md
├── 02-nav2-bringup-lifecycle-actions.md
├── 03-nav2-bt-navigator-and-bt-xml.md
├── 04-nav2-costmaps-and-layers.md
├── 05-nav2-global-planning.md
├── 06-nav2-local-control-and-cmdvel.md
├── 07-nav2-localization-odom-amcl-ekf.md
├── 08-nav2-recoveries-progress-and-goal-checkers.md
├── 09-nav2-waypoints-docking-zones-and-missions.md
├── 10-nav2-parameters-launch-and-plugin-extension.md
├── 11-nav2-debugging-observability-and-bag-analysis.md
├── 12-nav2-amr-failure-patterns-and-capstone.md
├── 13-nav2-senior-interview-questions.md
└── exercises/
    ├── 01-bringup-and-lifecycle.md
    ├── 02-bt-trace-and-recovery-loop.md
    ├── 03-costmap-tuning-lab.md
    ├── 04-planner-comparison-lab.md
    ├── 05-controller-tuning-lab.md
    ├── 06-localization-contract-diagnosis.md
    ├── 07-recovery-policy-design.md
    ├── 08-zones-waypoints-and-missions.md
    ├── 09-plugin-extension-lab.md
    └── 10-amr-debug-capstone.md

These lesson and exercise files are the planned build-out for this track. This page is the roadmap for creating them in order and then studying them.

Time Budget

• Recommended duration: 8 to 10 weeks
• Suggested load: 4 to 6 hrs/week
• Total depth target: 40 to 55 hours

This assumes you want both conceptual depth and enough repetition to diagnose Nav2 incidents under pressure.

Track Outcomes

By the end of the track, you should be able to:

• Explain every major Nav2 server and how they interact during NavigateToPose
• Diagnose failures by separating localization, costmap, planner, controller, and BT-policy problems
• Tune a controller and costmap for an AMR in narrow aisles without relying on cargo-cult defaults
• Understand how AMCL, robot localization, wheel odometry, and IMU quality affect Nav2 behavior
• Design recovery policies that match real AMR operating conditions
• Add operational map semantics such as keepout zones, speed zones, waypoint flows, and docking behaviors
• Extend Nav2 with parameters, launch structure, or custom plugins when defaults are insufficient

Weekly Schedule

Week 1: System Architecture and Goal Flow

Planned lesson: 01 — Nav2 System Architecture

Focus:

• End-to-end flow of a goal through bt_navigator, planner, controller, behavior server, and lifecycle manager
• What Nav2 owns vs what localization, perception, and the rest of the AMR stack own
• How actions, topics, transforms, and parameters meet at runtime

Hands-on:

• Draw the node graph for a working Nav2 bringup from memory
• Trace one NavigateToPose request from UI or mission layer to /cmd_vel

Planned exercise:

• 01 — Bringup And Lifecycle

Senior interview focus:

• Explain the full navigation data flow without hand-waving
• Separate navigation orchestration from localization and perception responsibilities

Week 2: Bringup, Lifecycle, and Action Contracts

Planned lesson: 02 — Nav2 Bringup Lifecycle Actions

Focus:

• Lifecycle transitions and why partial bringup causes misleading runtime failures
• Action interfaces for NavigateToPose, NavigateThroughPoses, recovery behaviors, and waypoint flows
• Startup sequencing and server dependency ordering

Hands-on:

• Bring up Nav2 and intentionally break one dependency
• Observe lifecycle states and recover to a clean running system

Planned exercises:

• 01 — Bringup And Lifecycle
• 02 — Bt Trace And Recovery Loop

Senior interview focus:

• Explain lifecycle failure modes and what to inspect first in a broken bringup

Week 3: BT Navigator and BehaviorTree XML

Planned lesson: 03 — Nav2 Bt Navigator And Bt Xml

Focus:

• PipelineSequence, RecoveryNode, ReactiveFallback, decorators, blackboard values, replanning cadence
• Default tree behavior vs AMR-customized tree behavior
• When a problem belongs in BT design vs parameters vs a plugin

Hands-on:

• Walk a default tree tick by tick for success, planner failure, and controller failure
• Modify a recovery ordering for a blocked aisle scenario

Planned exercises:

• 02 — Bt Trace And Recovery Loop
• 07 — Recovery Policy Design

Senior interview focus:

• Defend a custom BT recovery design for a warehouse robot under dynamic obstruction

Week 4: Costmaps, Layers, and Map Semantics

Planned lesson: 04 — Nav2 Costmaps And Layers

Focus:

• Global vs local costmaps, static layer, obstacle layer, inflation, footprint, rolling window, observation sources
• Phantom obstacles, stale obstacles, inflated corridors, unknown space policy
• Keepout zones and speed zones for AMR operations

Hands-on:

• Tune inflation radius and footprint on the same map and compare passability
• Inject a stale obstacle or bad observation source and inspect the resulting costmap

Planned exercises:

• 03 — Costmap Tuning Lab
• 08 — Zones Waypoints And Missions

Senior interview focus:

• Explain why a planner says “no path” in an apparently open corridor

Week 5: Global Planning

Planned lesson: 05 — Nav2 Global Planning

Focus:

• NavFn vs Smac 2D vs Smac Hybrid-A*
• Kinematic feasibility, unknown space, path quality, compute budget, aisle geometry
• Planner behavior under warehouse-style maps and blocked goals

Hands-on:

• Compare planner outputs on the same map under different footprint and costmap settings
• Explain when a planner failure is really a costmap failure upstream

Planned exercise:

• 04 — Planner Comparison Lab

Senior interview focus:

• Choose a planner for a differential-drive AMR in a constrained warehouse and defend the choice

Week 6: Local Control, Velocity Shaping, and Motion Quality

Planned lesson: 06 — Nav2 Local Control And Cmdvel

Focus:

• DWB and Regulated Pure Pursuit
• Critics, lookahead, goal tolerances, progress checkers, goal checkers, velocity smoother
• Oscillation, corner cutting, sluggish approach, and over-conservative slowdown

Hands-on:

• Tune controller behavior on tight turns, aisle entry, and docking approach
• Compare controller behavior before and after parameter changes

Planned exercise:

• 05 — Controller Tuning Lab

Senior interview focus:

• Diagnose whether a motion-quality issue is controller tuning, bad path quality, or bad localization

Week 7: Localization Contracts and TF Correctness for Nav2

Planned lesson: 07 — Nav2 Localization Odom Amcl Ekf

Focus:

• What Nav2 expects from map, odom, and base_link
• AMCL vs robot_localization, wheel odometry, IMU fusion, transform tolerances, and timestamp drift
• Why localization defects masquerade as planning or controller problems

Hands-on:

• Replay a bag with localization drift and identify downstream Nav2 symptoms
• Diagnose TF lookup failures and stale transform timing

Planned exercise:

• 06 — Localization Contract Diagnosis

Senior interview focus:

• Explain what breaks when odom is locally smooth but globally wrong

Week 8: Recoveries, Goal Checking, and Mission Robustness

Planned lesson: 08 — Nav2 Recoveries Progress And Goal Checkers

Focus:

• Clear costmaps, spin, backup, wait, retry budgets, progress checker, goal checker
• What recoveries should exist for blocked aisles, temporary humans, phantom obstacles, and stuck robots
• How AMR missions should degrade instead of failing opaquely

Hands-on:

• Design a recovery chain for three AMR incident patterns
• Explain when to stop retrying and escalate to the higher-level mission system

Planned exercise:

• 07 — Recovery Policy Design

Senior interview focus:

• Design a recovery policy that is safe, observable, and operationally useful

Week 9: Waypoints, Docking, Zones, and AMR Mission Features

Planned lesson: 09 — Nav2 Waypoints Docking Zones And Missions

Focus:

• Waypoint follower, route execution, docking, semantic map overlays, no-go areas, speed zones
• Mission-layer interactions with Nav2 and when mission bugs look like navigation bugs

Hands-on:

• Model a warehouse lane policy with zones and speed constraints
• Walk a multi-stop route and reason about what belongs in Nav2 vs the task layer

Planned exercise:

• 08 — Zones Waypoints And Missions

Senior interview focus:

• Explain how you would represent keepout zones, preferred lanes, and docking constraints cleanly

Week 10: Extension, Observability, and AMR Capstone

Planned lessons:

• 10 — Nav2 Parameters Launch And Plugin Extension
• 11 — Nav2 Debugging Observability And Bag Analysis
• 12 — Nav2 Amr Failure Patterns And Capstone
• 13 — Nav2 Senior Interview Questions

Focus:

• Parameter layering, launch composition, namespaces, plugin choices, and when to write custom code
• Bag-based diagnosis, topic/TF inspection, action tracing, costmap introspection, and failure pattern triage
• Production scenarios: narrow-aisle aborts, repeated recovery loops, phantom obstacles, stale localization, dynamic pallet intrusion

Hands-on:

• Run a full capstone RCA from logs, topics, TF, and parameters
• Propose the smallest fix that addresses the real root cause

Planned exercises:

• 09 — Plugin Extension Lab
• 10 — Amr Debug Capstone

Senior interview focus:

• Debug a field incident with incomplete evidence and justify the first three things you would inspect

Exercise Model

Every module in this track should eventually include three exercise types:

1. Inspection drill Example: read a BT snippet, costmap config, or topic list and explain what it will do.

2. Hands-on lab Example: run Nav2, tune a parameter, replay a bag, or inspect lifecycle state transitions.

3. AMR diagnosis task Example: use logs and symptoms to determine whether the root cause lives in localization, costmaps, planning, control, or BT logic.

The capstone should combine all three.

Senior-Level Interview Categories

The final interview bank should be grouped by category, not by random questions.

1. System Architecture and Goal Flow

• Walk through NavigateToPose end to end
• Explain which node owns which responsibility and what inputs each requires

2. TF, Time, and Localization Correctness

• Explain map, odom, and base_link semantics
• Diagnose transform tolerance and timestamp drift issues

3. Costmaps and Perception Integration

• Explain layer composition, stale obstacles, inflation side effects, and observation-source bugs

4. Planner Selection and Tradeoffs

• Compare NavFn, Smac 2D, and Hybrid-A* for real AMR constraints

5. Controller Tuning and Motion Quality

• Diagnose oscillation, slowdown, corner cutting, and poor goal convergence

6. BT Design and Recovery Policy

• Decide when a fix belongs in BT XML, parameters, or a plugin

7. Production Debugging

• Explain how to separate localization faults from planning or control faults quickly

8. AMR Product Constraints

• Handle speed zones, keepout zones, docking, waypoint missions, and dynamic warehouse traffic

9. Plugin Extension and Code Ownership

• Decide when to write a custom BT node, planner wrapper, or controller plugin

10. Senior Judgment

• Decide what to instrument first during a real navigation incident
• Choose defaults for a new AMR platform and defend the tradeoffs

Suggested Hands-On Environment

Use whichever environment is available, but the track should be designed to work with one of these setups:

• Gazebo or Isaac Sim with a differential-drive warehouse robot
• Recorded AMR bags with TF, odom, scan, costmaps, and Nav2 action traces
• A local Nav2 bringup with maps, lifecycle nodes, and CLI tools

The exercises should not depend on owning hardware. Hardware-specific tuning can be an optional extension.

Completion Standard

You are done with this track when you can do all of the following without checking references:

• Draw the Nav2 node and server architecture from memory
• Explain why a recovery loop is happening from logs alone
• Read a costmap configuration and predict likely runtime behavior
• Diagnose whether a failure is localization, costmap, planner, controller, or BT policy
• Propose a safe AMR-specific Nav2 tuning change and defend it technically
• Answer senior-level Nav2 interview questions without drifting into generic ROS 2 answers

Next Step

Start here after finishing the base ROS 2 track by using this roadmap page, then writing or opening the first planned lesson: 01 — Nav2 System Architecture

Recommended first reading order inside this subtrack:

1. 01 — Nav2 System Architecture
2. 02 — Nav2 Bringup Lifecycle Actions
3. 03 — Nav2 Bt Navigator And Bt Xml

Once those are comfortable, move into costmaps, planning, control, and AMR-specific debugging.