{
  "schemaVersion": 1,
  "id": "atlas-lesson-robot-goal-plan-action",
  "title": "Goals, plans and actions are different objects",
  "category": "learning",
  "format": "Atlas lesson",
  "description": "Goal constraints, collision-aware planning, timed execution, feedback and cancellation.",
  "canonicalUrl": "https://semiaifoundry.com/learn/atlas/robot-goal-plan-action/",
  "publishedDate": null,
  "revisionDate": "2026-10-07",
  "author": null,
  "topic": "Observations become physical actions",
  "series": null,
  "license": null,
  "rightsUrl": "https://semiaifoundry.com/ai-atlas/licenses.html",
  "version": "1.3.0",
  "status": "Published educational lesson",
  "evidenceScope": "This is a worked explanation, not an executing external platform, model or robot. The quantities and fixtures are authored. Source scope is recorded in the automation learning review manifest; no source benchmark or private system behavior is claimed. A kinematic lower bound with no acceleration or stop-time model. It assumes the named path is feasible; it does not certify collision avoidance or physical safety.",
  "contentCoverage": "published-lesson",
  "artifactUrl": "https://semiaifoundry.com/ai-atlas/lesson-data/robot-goal-plan-action-38ee541b0248a399.json",
  "interactiveUrl": "https://semiaifoundry.com/ai-atlas/#robot-goal-plan-action",
  "related": [],
  "supersedes": [],
  "supersededBy": [],
  "contentHash": "2f2baf0df2fdcaf125ffda4dd14b5313b25a1998e3830edb2c07ecd2e22b17aa",
  "contentLength": 6168,
  "jsonUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-robot-goal-plan-action.json",
  "markdownUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-robot-goal-plan-action.md",
  "sections": [
    {
      "id": "section-1",
      "title": "Goals, plans and actions are different objects",
      "start": 53,
      "end": 494
    },
    {
      "id": "section-2",
      "title": "The mechanism",
      "start": 494,
      "end": 1676
    },
    {
      "id": "section-3",
      "title": "Science and assumptions",
      "start": 1676,
      "end": 1704
    },
    {
      "id": "section-4",
      "title": "Goals, plans and actions are different objects · explicit relationship",
      "start": 1704,
      "end": 2703
    },
    {
      "id": "section-5",
      "title": "Engineering example",
      "start": 2703,
      "end": 3229
    },
    {
      "id": "section-6",
      "title": "Distinctions and tradeoffs",
      "start": 3229,
      "end": 3576
    },
    {
      "id": "section-7",
      "title": "Specifications and interfaces",
      "start": 3576,
      "end": 4041
    },
    {
      "id": "section-8",
      "title": "Interfaces",
      "start": 4041,
      "end": 4317
    },
    {
      "id": "section-9",
      "title": "Check your understanding",
      "start": 4317,
      "end": 4346
    },
    {
      "id": "section-10",
      "title": "A 2.5 m path has a 0.5 m/s speed limit. What is its ideal minimum traversal time?",
      "start": 4346,
      "end": 4655
    },
    {
      "id": "section-11",
      "title": "Evidence and scope",
      "start": 4655,
      "end": 5230
    },
    {
      "id": "section-12",
      "title": "Sources",
      "start": 5230,
      "end": 5707
    },
    {
      "id": "section-13",
      "title": "Continue exploring",
      "start": 5707,
      "end": 5730
    },
    {
      "id": "section-14",
      "title": "Foundations",
      "start": 5730,
      "end": 5978
    },
    {
      "id": "section-15",
      "title": "Related lessons",
      "start": 5978,
      "end": 6168
    }
  ],
  "citations": [
    {
      "title": "Explore in Atlas",
      "url": "https://semiaifoundry.com/ai-atlas/#robot-goal-plan-action"
    },
    {
      "title": "ROS 2 Jazzy: Understanding actions",
      "url": "https://raw.githubusercontent.com/ros2/ros2_documentation/jazzy/source/Concepts/Basic/About-Actions.rst"
    },
    {
      "title": "MoveIt Planning Scene",
      "url": "https://moveit.picknik.ai/main/doc/examples/planning_scene/planning_scene_tutorial.html"
    },
    {
      "title": "Published lesson data",
      "url": "https://semiaifoundry.com/ai-atlas/lesson-data/robot-goal-plan-action-38ee541b0248a399.json"
    },
    {
      "title": "Atlas releases and licensing",
      "url": "https://semiaifoundry.com/ai-atlas/licenses.html"
    },
    {
      "title": "Robot observations have coordinates and age",
      "url": "https://semiaifoundry.com/learn/atlas/robot-observation-frame/"
    },
    {
      "title": "Human review, interruption and cancellation",
      "url": "https://semiaifoundry.com/learn/atlas/human-review-cancellation/"
    },
    {
      "title": "Tools, memory & agent loop",
      "url": "https://semiaifoundry.com/learn/atlas/tool-loop/"
    },
    {
      "title": "Evaluation & the AGI frontier",
      "url": "https://semiaifoundry.com/learn/atlas/evaluation/"
    }
  ],
  "recordHash": "6b83570b49886a541d89d2cb7285f0b6f82a6e58897e50eeb9bbbc937fb66b9a",
  "contentFormat": "text/markdown",
  "content": "Observations become physical actions · Atlas v1.3.0\n\n# Goals, plans and actions are different objects\n\nGoal constraints, collision-aware planning, timed execution, feedback and cancellation.\n\n[Explore in Atlas](https://semiaifoundry.com/ai-atlas/#robot-goal-plan-action)Logical state, task evidence and declared units; no physical assembly scale\n\nLearning objectiveDistinguish accepted goals, feasible plans, executed motion and observed completion under explicit geometric and timing limits.\n\n## The mechanism\n\nA goal states a desired outcome or constraint. A plan describes a route or trajectory intended to reach it. An action executes over time and returns progress or a terminal outcome. These objects can fail at different boundaries: a goal may be unreachable, a path may collide, an initially valid path may become stale, or execution may fail to follow it. ROS 2 actions provide a long-running goal/feedback/cancellation interface, while a planning scene supplies robot and environment state for collision and constraint checks. Neither an accepted goal nor a successful planning call proves physical completion. The worked path bound below is deliberately simple. No motion planner or robot actuator runs in this lesson, and the bound is not a safety certificate.\n\n1.  The goal names a target and its acceptable tolerance in a declared frame.\n2.  Planning uses the current scene to find a path satisfying geometric and kinematic constraints.\n3.  Execution follows a timed command and reports observed progress, not merely the planned coordinates.\n4.  A changed obstacle or cancellation request is handled as a new event; completion is checked against actual state.\n\n## Science and assumptions\n\n### Goals, plans and actions are different objects · explicit relationship\n\n$$\nT\\ge\\frac{L}{v_{\\max}}\n$$\n\nTraversal time is at least path length divided by the maximum permitted speed.\n\nSymbols and units\n\nT\n\nTraversal duration\n\ns\n\nL\n\nLength of the chosen feasible path\n\nm\n\nvmax\n\nPositive upper bound on path speed\n\nm/s\n\nA cart’s collision-free path is 2 m long and its speed is bounded by 0.5 m/s. Even with instantaneous acceleration and no pauses, traversal needs at least 4 s. A new obstacle forces a 2.5 m detour, raising that lower bound to 5 s. A plan computed before the obstacle appeared cannot be treated as currently collision-free. If cancellation is requested after motion begins, a later stopped state must be observed; the request itself does not prove zero velocity or erase the distance already traveled.\n\nAssumptions A kinematic lower bound with no acceleration or stop-time model. It assumes the named path is feasible; it does not certify collision avoidance or physical safety.\n\n## Engineering example\n\nA cart’s collision-free path is 2 m long and its speed is bounded by 0.5 m/s. Even with instantaneous acceleration and no pauses, traversal needs at least 4 s. A new obstacle forces a 2.5 m detour, raising that lower bound to 5 s. A plan computed before the obstacle appeared cannot be treated as currently collision-free. If cancellation is requested after motion begins, a later stopped state must be observed; the request itself does not prove zero velocity or erase the distance already traveled.\n\n## Distinctions and tradeoffs\n\nPlanning success or action acceptance is not evidence that a robot reached the goal.\n\nReplanning can restore feasibility after scene changes but costs time and depends on sufficiently current observations.\n\nLanguage or learned policies need a distinct planning and action boundary before their goals become motion.\n\n## Specifications and interfaces\n\nMechanism boundary\n\nWhy can an accepted plan still fail to reach its goal?\n\nA kinematic lower bound with no acceleration or stop-time model. It assumes the named path is feasible; it does not certify collision avoidance or physical safety.\n\nEvidence type\n\nAuthored worked example\n\nWorked explanation with a bounded numeric or state example. This lesson does not claim an executing implementation of an external model or platform.\n\n### Interfaces\n\nDeclared input and state Input\n\nThe goal names a target and its acceptable tolerance in a declared frame.\n\nChecked outcome Output\n\nDistinguish accepted goals, feasible plans, executed motion and observed completion under explicit geometric and timing limits.\n\n## Check your understanding\n\n### A 2.5 m path has a 0.5 m/s speed limit. What is its ideal minimum traversal time?\n\n1.  1.25 s\n2.  5 s\n3.  No time is needed once the goal is accepted\n\nAnswer and reasoning\n\n5 s\n\nDistance divided by the maximum speed gives a lower bound; acceleration and delays can only add time under these assumptions.\n\n## Evidence and scope\n\nThis is a worked explanation, not an executing external platform, model or robot. The quantities and fixtures are authored. Source scope is recorded in the automation learning review manifest; no source benchmark or private system behavior is claimed. A kinematic lower bound with no acceleration or stop-time model. It assumes the named path is feasible; it does not certify collision avoidance or physical safety.\n\nThe worked example explains this mechanism. General numerical labs are a reference; they do not execute this lesson’s full mechanism.\n\n## Sources\n\n-   [ROS 2 Jazzy: Understanding actions](https://raw.githubusercontent.com/ros2/ros2_documentation/jazzy/source/Concepts/Basic/About-Actions.rst)\n-   [MoveIt Planning Scene](https://moveit.picknik.ai/main/doc/examples/planning_scene/planning_scene_tutorial.html)\n\n[Published lesson data](https://semiaifoundry.com/ai-atlas/lesson-data/robot-goal-plan-action-38ee541b0248a399.json) · [Atlas releases and licensing](https://semiaifoundry.com/ai-atlas/licenses.html)\n\n## Continue exploring\n\n### Foundations\n\n-   [Robot observations have coordinates and age](https://semiaifoundry.com/learn/atlas/robot-observation-frame/)\n-   [Human review, interruption and cancellation](https://semiaifoundry.com/learn/atlas/human-review-cancellation/)\n\n### Related lessons\n\n-   [Tools, memory & agent loop](https://semiaifoundry.com/learn/atlas/tool-loop/)\n-   [Evaluation & the AGI frontier](https://semiaifoundry.com/learn/atlas/evaluation/)"
}