Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fast End-to-End Generation of Belief Space Paths for Minimum Sensing Navigation

Lukas Taus, Vrushabh Zinage, Takashi Tanaka, Richard Tsai

TL;DR

An approach that leverages a deep learning model to predict optimal path candidates directly from the problem description and significantly reduces computation time compared to the sampling-based baseline algorithm is proposed.

Abstract

We revisit the problem of motion planning in the Gaussian belief space. Motivated by the fact that most existing sampling-based planners suffer from high computational costs due to the high-dimensional nature of the problem, we propose an approach that leverages a deep learning model to predict optimal path candidates directly from the problem description. Our proposed approach consists of three steps. First, we prepare a training dataset comprising a large number of input-output pairs: the input image encodes the problem to be solved (e.g., start states, goal states, and obstacle locations), whereas the output image encodes the solution (i.e., the ground truth of the shortest path). Any existing planner can be used to generate this training dataset. Next, we leverage the U-Net architecture to learn the dependencies between the input and output data. Finally, a trained U-Net model is applied to a new problem encoded as an input image. From the U-Net's output image, which is interpreted as a distribution of paths,an optimal path candidate is reconstructed. The proposed method significantly reduces computation time compared to the sampling-based baseline algorithm.

Fast End-to-End Generation of Belief Space Paths for Minimum Sensing Navigation

TL;DR

An approach that leverages a deep learning model to predict optimal path candidates directly from the problem description and significantly reduces computation time compared to the sampling-based baseline algorithm is proposed.

Abstract

We revisit the problem of motion planning in the Gaussian belief space. Motivated by the fact that most existing sampling-based planners suffer from high computational costs due to the high-dimensional nature of the problem, we propose an approach that leverages a deep learning model to predict optimal path candidates directly from the problem description. Our proposed approach consists of three steps. First, we prepare a training dataset comprising a large number of input-output pairs: the input image encodes the problem to be solved (e.g., start states, goal states, and obstacle locations), whereas the output image encodes the solution (i.e., the ground truth of the shortest path). Any existing planner can be used to generate this training dataset. Next, we leverage the U-Net architecture to learn the dependencies between the input and output data. Finally, a trained U-Net model is applied to a new problem encoded as an input image. From the U-Net's output image, which is interpreted as a distribution of paths,an optimal path candidate is reconstructed. The proposed method significantly reduces computation time compared to the sampling-based baseline algorithm.
Paper Structure (13 sections, 8 equations, 11 figures)

This paper contains 13 sections, 8 equations, 11 figures.

Table of Contents

  1. Introduction
  2. Problem Statement
  3. The Proposed Approach
  4. Data Collection
  5. Path Prediction
  6. Path Reconstruction Algorithm
  7. Results
  8. Experiment Setup
  9. Generalizability
  10. Path Quality
  11. Computational Efficiency
  12. Conclusion
  13. Acknowledgements

Figures (11)

  • Figure 1: The RI-RRT* algorithm applied to Problem 1.
  • Figure 2: Our approach replaces the traditional algorithm (RI-RRT*) shown in red with the three steps in blue.
  • Figure 3: Visual representation of the encoding scheme for available information and the computed optimal path
  • Figure 4: Full architecture of a UNet. The diagram illustrates the sequence of encoding and decoding blocks, skip connections, and the bottleneck layer. Each block represents a distinct operation, with arrows indicating data flow.
  • Figure 5: The left column shows the probability density predicted by the neural network and the right column shows the reference solution computed by RI-RRT* for the same obstacle map, start, and target locations.
  • ...and 6 more figures