Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Diffusion Model for Planning: A Systematic Literature Review

Toshihide Ubukata, Jialong Li, Kenji Tei

TL;DR

This systematic review maps diffusion-model-based planning across robotics, autonomous driving, and instructional tasks, highlighting how diffusion processes enable robust, iterative trajectory generation and policy scheduling. It organizes the literature into foundational models, motion/path planning, skill learning, safety/uncertainty, and domain-specific applications, identifying key methods such as Diffuser, Diffusion-QL, and AlignDiff, as well as practical datasets and benchmarks. The analysis uncovers advances in efficiency (latent action spaces, equivariant diffusion), conditional planning (context, skills, and offline meta-RL), and safety mechanisms (control barrier functions, restoration gap, LTL constraints), while also outlining challenges in scalability, real-time operation, generalization, and human-robot interaction. Overall, diffusion-based planning offers flexible, data-efficient tools for complex, long-horizon decision-making with promising implications for real-world autonomy and robotics.

Abstract

Diffusion models, which leverage stochastic processes to capture complex data distributions effectively, have shown their performance as generative models, achieving notable success in image-related tasks through iterative denoising processes. Recently, diffusion models have been further applied and show their strong abilities in planning tasks, leading to a significant growth in related publications since 2023. To help researchers better understand the field and promote the development of the field, we conduct a systematic literature review of recent advancements in the application of diffusion models for planning. Specifically, this paper categorizes and discusses the current literature from the following perspectives: (i) relevant datasets and benchmarks used for evaluating diffusion modelbased planning; (ii) fundamental studies that address aspects such as sampling efficiency; (iii) skill-centric and condition-guided planning for enhancing adaptability; (iv) safety and uncertainty managing mechanism for enhancing safety and robustness; and (v) domain-specific application such as autonomous driving. Finally, given the above literature review, we further discuss the challenges and future directions in this field.

Diffusion Model for Planning: A Systematic Literature Review

TL;DR

This systematic review maps diffusion-model-based planning across robotics, autonomous driving, and instructional tasks, highlighting how diffusion processes enable robust, iterative trajectory generation and policy scheduling. It organizes the literature into foundational models, motion/path planning, skill learning, safety/uncertainty, and domain-specific applications, identifying key methods such as Diffuser, Diffusion-QL, and AlignDiff, as well as practical datasets and benchmarks. The analysis uncovers advances in efficiency (latent action spaces, equivariant diffusion), conditional planning (context, skills, and offline meta-RL), and safety mechanisms (control barrier functions, restoration gap, LTL constraints), while also outlining challenges in scalability, real-time operation, generalization, and human-robot interaction. Overall, diffusion-based planning offers flexible, data-efficient tools for complex, long-horizon decision-making with promising implications for real-world autonomy and robotics.

Abstract

Diffusion models, which leverage stochastic processes to capture complex data distributions effectively, have shown their performance as generative models, achieving notable success in image-related tasks through iterative denoising processes. Recently, diffusion models have been further applied and show their strong abilities in planning tasks, leading to a significant growth in related publications since 2023. To help researchers better understand the field and promote the development of the field, we conduct a systematic literature review of recent advancements in the application of diffusion models for planning. Specifically, this paper categorizes and discusses the current literature from the following perspectives: (i) relevant datasets and benchmarks used for evaluating diffusion modelbased planning; (ii) fundamental studies that address aspects such as sampling efficiency; (iii) skill-centric and condition-guided planning for enhancing adaptability; (iv) safety and uncertainty managing mechanism for enhancing safety and robustness; and (v) domain-specific application such as autonomous driving. Finally, given the above literature review, we further discuss the challenges and future directions in this field.
Paper Structure (31 sections, 2 figures, 4 tables)

This paper contains 31 sections, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Literature Search Methodology
  3. Dataset and Benchmark
  4. Motion and Path Planning
  5. 3D Planning
  6. Instructional Video
  7. Autonomous Driving
  8. Safety-Critical Datasets
  9. Fundamental and General Planning
  10. Foundation of Diffusion Model for Planning
  11. Motion and Path Planning
  12. Efficiency
  13. Preference Customization
  14. Replanning
  15. Skill Learning and Conditional Planning
  16. ...and 16 more sections

Figures (2)

  • Figure 1: Trends in paper numbers searched using keywords "Diffusion" and "Plan" in ArXiv.
  • Figure 2: Overview of diffusion model category for planning