KGLAMP: Knowledge Graph-guided Language model for Adaptive Multi-robot Planning and Replanning

TL;DR

KGLAMP introduces a knowledge-graph-grounded LLM planning framework for heterogeneous multi-robot teams, grounding planning in a union of relation, property, and reach graphs to produce accurate PDDL problem definitions. A replanning loop updates the knowledge graphs via failure diagnosis and VLM-assisted discovery under partial observability, enabling robust adaptation to dynamic environments. Empirical results on MAT-THOR show substantial improvements in task completion and robustness over LLM-only and PDDL-based baselines, highlighting the value of explicit grounded representations and memory for long-horizon planning. The work suggests practical benefits for real-world deployments and outlines future directions in perception integration and model distillation to improve scalability and efficiency.

Abstract

Heterogeneous multi-robot systems are increasingly deployed in long-horizon missions that require coordination among robots with diverse capabilities. However, existing planning approaches struggle to construct accurate symbolic representations and maintain plan consistency in dynamic environments. Classical PDDL planners require manually crafted symbolic models, while LLM-based planners often ignore agent heterogeneity and environmental uncertainty. We introduce KGLAMP, a knowledge-graph-guided LLM planning framework for heterogeneous multi-robot teams. The framework maintains a structured knowledge graph encoding object relations, spatial reachability, and robot capabilities, which guides the LLM in generating accurate PDDL problem specifications. The knowledge graph serves as a persistent, dynamically updated memory that incorporates new observations and triggers replanning upon detecting inconsistencies, enabling symbolic plans to adapt to evolving world states. Experiments on the MAT-THOR benchmark show that KGLAMP improves performance by at least 25.5% over both LLM-only and PDDL-based variants.

Figures (7)

• Figure 1: Impact of relational knowledge on task planning. (a) Without relational graphs, PDDL models fail to capture object relationships, leading to infeasible actions. (b) Integrating relationship, property, and reachability graphs enables accurate PDDL generation and feasible task plans.
• Figure 2: Minimal STRIPS PDDL example illustrating (a) Domain PDDL and (b) Problem PDDL
• Figure 3: Overview of KGLAMP framework. Environment and robot information are encoded as relationship, property, and reachability knowledge graphs. LLM agents generate goal, relational, property, and reachability predicates in a dependency-aware manner to synthesize a PDDL problem, execute the resulting plan, and iteratively update the graphs and replan upon execution failures.
• Figure 4: An example knowledge graph. (a) $G_{\text{relation}}$ captures semantic and geometric relationships among objects. (b) $G_{\text{property}}$ encodes object attributes and robot capabilities. (c) $G_{\text{reach}}$ models spatial connectivity.
• Figure 5: An example of LLM prompt for $LLM_\text{relation}$. This prompt utilizes contextual examples, scenario definition, spatial data, and output constraints to extract relevant spatial tuples.