cs.AIAug 20, 2026

Towards general embodied intelligence: integrating large language models, knowledge bases, and reasoning capabilities to build the next generation of AI agents

Authors: Fujiang Yuan, Xia Huang, Lusheng Wang, Jun Ding, Zhen Tian, Yuxin Wang, Shaojie Gu, Yuki Funabora, 

Organizations: College of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, China · James Watt School of Engineering, University of Glasgow, G12 8QQ, UK · School of Energy and Power, Jiangsu University of Science and Technology, Zhenjiang 212100, China · Magnesium Research Center, Kumamoto University, Kumamoto 860-8555, Japan · 

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

The convergence of large language models (LLMs), structured knowledge bases (KBs), and reasoning ability (RA) presents a promising trajectory toward general embodied intelligence (GEI). This paper reviews the evolution of LLM-centered intelligent systems, emphasising their integration with knowledge representation, logical reasoning, and physical embodiment. We analyse LLM architectures, pre-training methods, and inference mechanisms, along with their interaction with external knowledge sources and structured reasoning frameworks. Furthermore, we examine embodied intelligence (EI) paradigms wherein agents learn and act in physical environments. To synthesise these dimensions, we present a conceptual framework that illustrates the synergy among LLMs, KBs, RA, and embodiment, serving as a guiding model for perception, reasoning, and action rather than an implemented engineering architecture. To advance toward GEI, we identify five key challenges: efficient LLM deployment, closed-loop knowledge integration, hybrid symbolic-neural reasoning, perception-action grounding, and continual learning. This survey provides a comprehensive roadmap for developing adaptive, multimodal agents capable of operating in complex, dynamic settings.

Figures & tables

AuthorsYearCore ContributionsLimitations
Yao et al. [ 8 ]2025Reviews multimodal embodied intelligence tasks and models; explores LLM/MLM integration with robots; proposes PPA paradigm and evaluates MLMs.Classical AI lacks physical interaction; DRL is data-hungry and less robust; VLN training is costly and generalization is poor; LLMs lack causal reasoning; Sim2Real transfer is weak; evaluation standards are lacking.
Dong et al. [ 9 ]2025Analyzes LLM-enabled Human-Robot Symbiotic Manufacturing (HRSM), including interaction, collaboration, and execution.Privacy and data sensitivity risks; hallucination and bias in LLMs; absence of unified evaluation standards; integration difficulties in diverse environments; token length limits document processing.
Tan et al. [ 10 ]2025Proposes CEIM to fuse multi-source data and LLM reasoning for semantic integration and decision-making in customized manufacturing.Data fusion issues; low semantic alignment; real-time control accuracy is lacking; high computation cost; insufficient standardization of AI-device interfaces.
Jeong et al. [ 11 ]2024Surveys five directions of LLM/VLM in robotics: reward design, control, planning, manipulation, and scene understanding.Limited by embedded compute; weak sensor fusion; prompt bias and poor interpretability; lacks domain-specific knowledge; Sim2Real generalization is weak.
Wang et al. [ 12 ]2025Discusses LLMs in robotics tasks and evaluates GPT-4V; explores future applications in agriculture and medicine.Plan robustness is lacking; long prompts require expert input; fixed action sets reduce flexibility; API is closed-source; RGBD-based 3D reasoning is limited.
Our Survey2025Systematically analyzed the synergy among LLMs, knowledge bases, reasoning capabilities, and embodied intelligence; proposed a unified framework and five future directions for GEI.Limited empirical examples; insufficient discussion on multilingual settings and real-world deployment.
TABLE I: Comparative Summary of Related Embodied Intelligence Review Articles
Fig. 1: Schematic diagram of a unified multimodal large language model architecture.
Fig. 2: Evolution and Taxonomy of Large Language Models.
Fig. 3: Schematic architecture of the Transformer model.
Fig. 4: Construction Process of Large Language Models.
Fig. 5: Description of the execution process of the intelligent document processing and knowledge question answering system.
Author(s)Core ImprovementAdvantagesApplication Scenarios
Cummins et al. [ 40 ]Compiler optimization frameworkEnhances disassembly processesCompiler optimization
Hu et al. [ 41 ]Prompting strategies incorporating context to enhance VQA performanceStrengthens contextual understandingKnowledge-based applications
Hu et al. [ 42 ]LLM-TIKG: Threat Intelligence Knowledge Graph frameworkEnhances cybersecurity analysisThreat Intelligence Analysis
Qian et al. [ 43 ]Infrastructure for large-scale LLM trainingEfficiently handles large-scale training tasksLLM training infrastructure
Sun et al. [ 44 ]UniCoder: Intermediate representation to improve downstream NLP task processingEnhances the model’s ability to handle complex tasksModel scaling
Zhang et al. [ 45 ]BB-GeoGPT: Framework tailored for Geographic Information ScienceHighly adaptable to domain-specific knowledgeGeographic Information Science (GIS)
TABLE II: Comparison of Research on Large Language Models (LLMs)
Fig. 6: Flowchart of knowledge base construction and use in large language model.
Fig. 7: Summary of inference optimization techniques in large language models.
Fig. 8: The core components and development path of embodied intelligence.
Fig. 9: The development of embodied intelligence.
Fig. 10: Different categories of Embodied Robots: (a) Quadruped Robots, (b) Fixed-base Robots, (c) Tracked Robots, (d) Wheeled Robots, and (e) Humanoid Robots.
Fig. 11: Example diagram of the integration of embodied intelligence and large language model.
AuthorYearResearch TopicKey ContributionApplication Domain
Zardini et al. [ 101 ]2020Co-Design FrameworkStructured co-design for integrationGeneral systems
Ma et al. [ 102 ]2021Soft Robotics ActuatorsIntegrated sensing & actuationSoft robotics
Gupta et al. [ 104 ]2021DERL for AdaptationCo-evolution of body and controlLocomotion & control
Roy et al. [ 105 ]2021ML in Embodied SystemsEnergy-aware physical learningRobotics & ML
Mengaldo et al. [ 103 ]2022Physics ModelingPhysics-based modelingSoft robotics
Iida et al. [ 106 ]2022Temporal DynamicsTimescale-based self-organizationAdaptive systems
TABLE III: Summary of Key Contributions in Embodied Intelligence Research
Fig. 12: Embodied Intelligence, Large Language Model, Knowledge Base and Reasoning Ability together constitute the core architecture of the intelligent system.
Author(s)InnovationAdvantagesDrawbacks
Daniel et al. [ 111 ]Verifying RDF triples using LLMs by comparing RDF statements to external documentsHigh precision (88%) in RDF verificationRequires human oversight, method may be slower
Feng et al. [ 112 ]Knowledge Graph–based thought framework to improve LLM responses with verified KG informationReduces factual errors, enhances drug–cancer associationsLimited to biomedical, requires cross-domain testing
Dai et al. [ 113 ]Empirical study on LLMs’ comprehension of KG input formats and prompt organizationImproves LLM performance in KG-based tasksVarying effectiveness across different models, noisy subgraphs
Erickson et al. [ 114 ]Enhancing LLM workflows with RAG and KGs to ensure traceability and flexibilityMore accurate and explainable LLM outputsIntegration with external tools can be complex
Xiang et al. [ 115 ]LLM-based generative retrieval framework focusing on trustworthy search resultsImproves relevance, responsibility, and trustworthinessWorks better with certain query types, not universally applicable
Sequeda et al. [ 116 ]LLM-powered enterprise question answering, supported by KGsProvides formal validation and trusted data for queriesMay not scale well for large or complex queries in diverse enterprises
TABLE IV: LLM and Knowledge Graph Integration
AuthorsInnovationAdvantagesDrawbacks
Wang et al. [ 117 ]LLM-KGMQA: Multi-hop question-answering system for the medical field, addressing entity linking and reasoningImproved accuracy (99.80%) and robust performanceRequires significant computational resources for entity linking and multi-hop reasoning
Jiang et al. [ 118 ]Knowledge-guided agricultural LLM for domain-specific decision-makingState-of-the-art performance in agriculture, improves accuracy in agricultural queriesData scarcity in the domain, challenges in precision retrieval
Benjira et al. [ 119 ]LLM-augmented KG for SDG indicator mapping, integrating open dataHigh precision in mapping open data to SDGsRequires access to high-quality and diverse metadata
Freire et al. [ 120 ]LLM-powered knowledge-sharing system for manufacturing, retrieving information from factory documentationEnhances quick information retrieval and resolution of issuesPreference for human experts still prevails, open-source models may have privacy concerns
Xie et al. [ 121 ]Theme-based lecture summary system using LLMs and graph-based segmentationEnhances knowledge recall and student satisfactionEffectiveness varies with lecture structure
Sun et al. [ 122 ]SF-GPT: A training-free method for enhancing LLM-based KG constructionImproved recall and F1 score, no training requiredMay not work well with all types of datasets or knowledge fusion
TABLE V: LLM for Domain-Specific Applications
AuthorsYearInnovationAdvantagesLimitations
Wu et al. [ 123 ]2025Proposed Long-to-Short (L2S) reasoning and model merging.Increases reasoning efficiency while maintaining performance.Merging efficiency depends on model size, adds computational complexity.
Tung et al. [ 124 ]2025Introduced GreenMind, using strategy optimization for reasoning.Outperforms previous works on Vietnamese reasoning tasks.Limited to specific languages, struggles with low-resource languages.
Ma et al. [ 125 ]2025Introduced SQL-R1, a reinforcement learning-based model for NL2SQL.Improves accuracy in complex database tasks.Reinforcement learning training may suffer from cold start issues.
Costas [ 126 ]2024GNN-RAG combines GNN reasoning with LLMs in a retrieval-augmented generation style.Optimizes graph data processing, improves knowledge graph QA.Performance depends on graph data quality, limited in complex reasoning.
Sui et al. [ 127 ]2024Proposed FiDeLiS for factuality-enhanced LLM reasoning.Improves factuality and interpretability.Dependency on reasoning steps increases computational cost.
Luo et al. [ 128 ]2025Introduced RALLRec+ for enhancing recommendation system reasoning.Enhances recommendation reasoning, handles dynamic user preferences.Requires continuous data updates, may still be less effective in real-time applications.
TABLE VI: Category 1: Research Based on Reasoning Frameworks and Methods
AuthorsYearInnovationAdvantagesLimitations
Yang et al. [ 129 ]2025Introduced Reasoning-based Bias Detector (RBD).Reduces bias in evaluations, improves consistency.Needs significant supervision for bias detection.
Zhang et al. [ 130 ]2024Designed step-wise multimodal models for document reasoning.Enhances multimodal understanding, improves complex question answering.Relies heavily on high-quality labeled data.
Beary et al. [ 131 ]2025Introduced hybrid fine-tuning to enhance medical domain reasoning.Improves reasoning accuracy in radiation oncology.Domain-specific, limited transferability to other medical fields.
Tan et al. [ 132 ]2025Proposed Hydra framework for cross-source enhanced reasoning.Excels in multi-hop and multi-entity reasoning tasks.Requires diverse data sources, computationally intensive.
Kunat et al. [ 133 ]2025Merged language-specific LLMs with advanced reasoning models.Enhances reasoning in low-resource languages.Can be computationally expensive for merging models.
Lan et al. [ 134 ]2024Proposed LLM4QA for graph-based reasoning with SPARQL queries.Improves knowledge graph reasoning efficiency.Relies on structured knowledge graphs, limited in unstructured data.
TABLE VII: Category 2: Application-Specific Reasoning Research
AuthorsYearInnovationAdvantagesLimitations
Fan et al. [ 138 ]2025Proposed a three-stage framework for industrial robot task execution using LLMsOutstanding GPT-4 performance in industrial tasksLimited ability in 3D spatial planning
Chen et al. [ 139 ]2025Designed RoboGPT with long-term decomposition and feedback adjustmentAdaptive subgoal updatingRequires large datasets and complex modules
Ruaridh et al. [ 141 ]2025Introduced ELLMER with RAG for multi-step reasoningStrong sensorimotor adaptationTested mainly in home-like tasks
Li et al. [ 143 ]2024Proposed multi-modal LLMs for autonomous mining drivingApplicable to realistic mining scenariosConceptual work, lacks implementation
Song et al. [ 146 ]2023Introduced LLM-Planner for few-shot planningStrong performance with limited dataWeak environmental state modeling
Andrew et al. [ 147 ]2025Unified GEA model across embodied tasks via MLLMHigh cross-domain generalizationData-hungry and complex training
TABLE VIII: LLM-driven Embodied Task Execution and Planning
AuthorsYearInnovationAdvantagesLimitations
Sun et al. [ 140 ]2025Multi-level LLM framework for coal mine sensor analysisEfficient risk prediction and learningRequires simulated and expert knowledge
Abhay et al. [ 154 ]2024EnvGen: LLM-generated adaptive RL training environmentsBoosts training efficiencyDepends on feedback quality
Zheng et al. [ 155 ]2023Steve-Eye combines LLM and visual encodersEnables end-to-end multimodal interactionRequires large datasets
Ma et al. [ 156 ]2025Perspective on LLMs in autonomous drivingOffers future research outlookLacks technical solutions
Liu et al. [ 157 ]2024EAI-SIM simulation platform with ROSControls UAVs and arms in photo-realistic simNeeds powerful simulation hardware
Song et al. [ 158 ]2024Guide-LLM aids visually impaired via text mapsCombines commonsense and path planningSmall sample evaluation
TABLE IX: Multi-Modal Fusion and Perception Enhancement in Embodied Systems
AuthorsYearInnovationAdvantagesLimitations
Zhang et al. [ 162 ]2024ReAd mechanism improves multi-agent LLM efficiencyFewer LLM queries and agent stepsEvaluation mainly in simulation
Liu et al. [ 163 ]2024FaGeL uses smart fabric for implicit feedbackLearns user preference non-intrusivelyHigh system complexity
Guo et al. [ 164 ]2024Organizational prompting reduces LLM agent conflictBoosts teamwork and efficiencyValidation in real-world pending
TABLE X: Multi-Agent Collaboration and Organizational Strategies
AuthorsYearInnovationAdvantagesLimitations
Liu et al. [ 142 ]2025Contextual backdoor attacks in LLM-based agentsDual-modal triggers for covert actionsNo defense strategy yet in real use
Zhang et al. [ 144 ]2024Investigated LLM-based robot jailbreak threatsSystematic exposure of jailbreak scenariosNo concrete safety mechanisms
Li et al. [ 145 ]2024Universal benchmarking interface for LLM agentsFine-grained error taxonomyNo coverage of real multimodal scenes
Kovalev et al. [ 153 ]2022Reviewed LLM-based instruction planners without trainingWorks in zero-shot scenariosLack of real task validations
Szot et al. [ 165 ]2023LLaRP: RL-trained LLM policy with visionSuperior multi-task performanceNeeds intensive training and environment
Leon et al. [ 166 ]2024Studied virtual agent personality effects in VRExtrovert agents yield better experienceNo embodiment in physical robots
TABLE XI: Security and Robustness Challenges in Embodied LLMs
AuthorsYearInnovationAdvantagesLimitations
Ruaridh et al. [ 168 ]2025Proposed ELLMER framework combining GPT-4 with RAG for task planning in unpredictable environments.Enables long-horizon task execution via visual/force feedback; adaptable planning.Generalization scope not fully validated.
Fan et al. [ 169 ]2024Introduced LLM agents in manufacturing for autonomous design, planning, and control.Achieved 81.88% task completion with GPT-4 in complex scenarios.Challenges remain in 3D spatial task handling and real-time integration.
Liang et al. [ 170 ]2023Introduced “Code as Policies” using LLMs to write robot policy code from commands.Enables reactive and trajectory-based control via few-shot prompting.Limited physical-world validation.
Xu et al. [ 171 ]2025Reviewed LLM-integrated CPSS frameworks for human-centered Industry 5.0.Connects LLMs, perception, scheduling, and swarm intelligence into unified framework.Lacks empirical task-specific evaluations.
Zhao et al. [ 172 ]2024Developed MultiBotGPT system using GPT-3.5 for UAV/UGV task execution.Outperforms BERT in assignment success rate; improves operator experience.Focused on limited command types and single-modal input.
Shen et al. [ 173 ]2024Discussed LLMs as foundation for embodied agents in physical environments.Conceptual bridge from virtual to real-world through embodiment.Lacks concrete framework or task-based implementation details.
TABLE XII: Summary of Studies on LLM-driven Embodied Control
More in this topic

Embodied AI

Embodied AI studies agents that perceive, reason, and act through bodies in physical or simulated environments.

See all 255 papers

Explore similar work

CardsList
Jan 18, 2026

A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents

Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) demonstrate strong reasoning capabilities in closed-world settings, they struggle in open-ended and dynamic environments. Agentic reasoning marks a paradigm shift by reframing LLMs as autonomous agents that plan, act, and learn through continual interaction. In this survey, we organize agentic reasoning along three complementary dimensions. First, we characterize environmental dynamics through three layers: foundational agentic reasoning, which establishes core single-agent capabilities including planning, tool use, and search in stable environments; self-evolving agentic reasoning, which studies how agents refine these capabilities through feedback, memory, and adaptation; and collective multi-agent reasoning, which extends intelligence to collaborative settings involving coordination, knowledge sharing, and shared goals. Across these layers, we distinguish in-context reasoning, which scales test-time interaction through structured orchestration, from post-training reasoning, which optimizes behaviors via reinforcement learning and supervised fine-tuning. We further review representative agentic reasoning frameworks across real-world applications and benchmarks, including science, robotics, healthcare, autonomous research, and mathematics. This survey synthesizes agentic reasoning methods into a unified roadmap bridging thought and action, and outlines open challenges and future directions, including personalization, long-horizon interaction, world modeling, scalable multi-agent training, and governance for real-world deployment.

Tianxin Wei, Ting-Wei Li, Zhining Liu +26

University of Illinois Urbana-Champaign · Meta · UCSD +2

Continual Learning for LLM AgentsMulti-Agent ReasoningAgentic AI
AskPDF
Aug 25, 2025

LLM-based Agentic Reasoning Frameworks: A Survey from Methods to Scenarios

Recent advances in LLM-based agents highlight the importance of their reasoning frameworks, which guide the problem-solving process in diverse ways. This survey introduces a unified formal language to systematically categorize these frameworks at three compositional levels: single-agent, tool-based, and multi-agent methods. Following our taxonomy, we review key application scenarios across scientific discovery, healthcare, software engineering, society, economics, and general-purpose tasks. It also compares the distinct features and evaluation strategies of each category. Through our taxonomy and comparisons, our survey explores the designs and strengths of LLM-based agentic frameworks in different scenarios, reviewing the fast-paced development of complex agentic systems in the real world.

Bingxi Zhao, Lin Geng Foo, Ping Hu +3

Multi-Agent LLM SystemsLLM Agents
AskPDF
Sep 16, 2026

A Comprehensive Review of Generative Physical Artificial Intelligence

The integration of large-scale foundation models with physical embodiments has led to significant advancements in robotics known as Generative Physical Artificial Intelligence (GPAI). These agentic AI systems autonomously perceive, reason, and act in complex real-world situations. This survey comprehensively analyzes GPAI systems, focusing on their architectural foundations, current applications, and key limitations. We introduce a taxonomy of five distinct approaches: Robot Foundation Models (RFMs) for cross-platform skill transfer; Vision-Language Action (VLA) models for end-to-end multi-modal perception and control; Large Behavior Models (LBMs) for human-like movement generation; Diffusion Policy Models (DPMs) for diffusion model-based temporally coherent action generation; and World Foundation Models (WFMs) for physics-compliant simulation and data generation. We examine how these approaches complement each other: WFMs generate training data for VLAs and DPMs, RFMs enable cross-platform deployment of learned policies, while LBMs provide motion priors for natural behavior. Through examples across autonomous vehicles, industrial automation, healthcare robotics, and humanoid systems, we identify significant performance improvements and summarize promising research directions in data-efficient learning, sim-to-real transfer, edge-compatible architectures, and safety frameworks. These insights advance embodied AI for IoT-connected environments where intelligent agents interact with networked sensors, actuators, and edge devices.

Satyam Gaba, Krutiksinh Rana, Siva Sai +2

Qualcomm Research, San Diego, USA · University of the Cumberlands, KY, USA · Department of Electronics and Instrumentation, BITS-Pilani, Goa Campus, India +3

Robot Foundation ModelsDiffusion PolicyEmbodied AIVLMs for RoboticsEmbodied World Models
AskPDF