Reasoning and Planning

Language2LTL

Translates natural-language task constraints into LTL-style checks that reject unsafe, skipped, or out-of-order plans.

Tool Introduction

Core parameters, trigger timing, and visual before/after demo references.

Short Explanation

Use Language2LTL before execution to make sure a VLM-produced plan respects safety rules and required step order.

InputVLM action plan + safety LTL formulas

OutputValidation result and violation feedback

Trigger TimingTriggered on demand after the required input files and configuration are prepared.

RuntimeLocal CPU

BeforeVLM action plan + safety LTL formulas

Prepare the scene, image, video, sensor stream, prompt, or configuration expected by the original project.

AfterValidation result and violation feedback

Read the produced visualization, prediction, map, trajectory, mask, grasp pose, or other documented artifact.

Preset Example

A quick-run style example for the documentation page.

Inputtools/language2ltl/examples/proposed_plan.json

Promptalways(grab_chemical -> historically(wear_gloves))

ExpectedA compliance flag and a human-readable violation reason when a prerequisite step is missing.

Parameters And Output

Readable controls and the meaning of each returned artifact.

Parameter Explanation

proposed_step_sequencepathtools/language2ltl/examples/proposed_plan.json

Ordered candidate steps produced by the planner.

active_ltl_constraintspathtools/language2ltl/examples/constraints.json

Temporal-logic constraints active for this task.

validator_modeselectmock_rule_check_no_install

Selects the installed Language2LTL backend or a local mock checker.

Output Explanation

is_sop_compliant

Whether the proposed sequence satisfies the active constraints.

violation_reason

Explanation of the first detected safety or ordering violation.

validation_trace

Optional trace of automaton or rule-checking states.

How To Use

Official resources, deployment steps, academic context, citation, and source-reported benchmark numbers.

Resources

GitHubhttps://github.com/JasonXinyuLiu/Language2LTL Code Downloadhttps://github.com/JasonXinyuLiu/Language2LTL/archive/refs/heads/main.zip Paperhttps://arxiv.org/abs/2305.07766

Deployment Notes

Clone or download the official Language2LTL repository.
Install the parser, planner, and temporal-logic dependencies required by the source project.
Prepare a proposed step sequence and active constraint file under tools/language2ltl/examples/.
Run the validator and save compliance reports under tools/language2ltl/runs/.

Relative Path Example

python tools/language2ltl/run.py --plan tools/language2ltl/examples/proposed_plan.json --constraints tools/language2ltl/examples/constraints.json --output tools/language2ltl/runs/validation.json

Expected Result Shape

{
  "tool": "language2ltl",
  "status": "ok",
  "results": [
    {
      "label": "Natural language to temporal-logic validation",
      "score": 0.87,
      "output": "Validation result and violation feedback"
    }
  ],
  "timing": {
    "runtime": "The submitted spreadsheet describes the local rule checker as millisecond-level; the upstream paper reports translation accuracy rather than a universal runtime.",
    "device": "documented in source benchmark when available"
  },
  "artifacts": {
    "visualization": "tools/language2ltl/runs/visualization.png",
    "raw_predictions": "tools/language2ltl/runs/predictions.json"
  }
}

Paper figure

Academic Info

Paper identity and contribution summary.

TitleLanguage2LTL: Translating Natural Language to Linear Temporal Logic for Robot Specification

AuthorsAdd authors

VenueIROS 2023

ContributionGrounds natural-language task rules into formal temporal constraints so robot plans can be checked before execution.

Citation

@misc{language2ltl2023,
  title={Language2LTL: Translating Natural Language to Linear Temporal Logic for Robot Specification},
  author={Author},
  year={2023},
  note={IROS 2023},
  url={https://arxiv.org/abs/2305.07766}
}

Benchmark

Only compact, source-reported numbers are shown here.

Dataset	Metric	Value	Runtime	Source
NL2TL, GPT-3-assisted data test	T5-large I.O./word accuracy	97.52+/-0.65%	Upstream NL-to-temporal-logic translation benchmark	Upstream proxy: NL2TL / Language2LTL paper, Table 1
NL-to-STL full task, Circuit / Navigation / Office email	T5-large + GPT-3 AP-detect accuracy	95.13+/-1.42% / 95.03+/-1.20% / 96.73+/-1.03%	Upstream full translation benchmark	Upstream proxy: NL2TL / Language2LTL paper, Table 2
AP detection, Circuit / Navigation / Office email	AP-detect accuracy	98.84+/-0.41% / 99.03+/-0.53% / 100.00+/-0.00%	Upstream AP detection benchmark	Upstream proxy: NL2TL / Language2LTL paper, Table 3

Artifacts

Official repository link, upstream NL2TL paper tables, mock rule-check example, constraint JSON, and validation output from the submitted spreadsheet.

Demo Images

Visual references from the original tool. Click any image to inspect the original size.