Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fast Confidence-Aware Human Prediction via Hardware-accelerated Bayesian Inference for Safe Robot Navigation

Michael Lu, Minh Bui, Xubo Lyu, Mo Chen

TL;DR

This work presents a simple yet novel approach for confidence-aware prediction that treats future predictions as particles, which supports finer prediction time steps, yielding more granular trajectory forecasts and allows motion planners to respond effectively to subtle changes in human behavior.

Abstract

As robots increasingly integrate into everyday environments, ensuring their safe navigation around humans becomes imperative. Efficient and safe motion planning requires robots to account for human behavior, particularly in constrained spaces such as grocery stores or care homes, where interactions with multiple individuals are common. Prior research has employed Bayesian frameworks to model human rationality based on navigational intent, enabling the prediction of probabilistic trajectories for planning purposes. In this work, we present a simple yet novel approach for confidence-aware prediction that treats future predictions as particles. This framework is highly parallelized and accelerated on an graphics processing unit (GPU). As a result, this enables longer-term predictions at a frequency of 125 Hz and can be easily extended for multi-human predictions. Compared to existing methods, our implementation supports finer prediction time steps, yielding more granular trajectory forecasts. This enhanced resolution allows motion planners to respond effectively to subtle changes in human behavior. We validate our approach through real-world experiments, demonstrating a robot safely navigating among multiple humans with diverse navigational goals. Our results highlight the methods potential for robust and efficient human-robot coexistence in dynamic environments.

Fast Confidence-Aware Human Prediction via Hardware-accelerated Bayesian Inference for Safe Robot Navigation

TL;DR

This work presents a simple yet novel approach for confidence-aware prediction that treats future predictions as particles, which supports finer prediction time steps, yielding more granular trajectory forecasts and allows motion planners to respond effectively to subtle changes in human behavior.

Abstract

As robots increasingly integrate into everyday environments, ensuring their safe navigation around humans becomes imperative. Efficient and safe motion planning requires robots to account for human behavior, particularly in constrained spaces such as grocery stores or care homes, where interactions with multiple individuals are common. Prior research has employed Bayesian frameworks to model human rationality based on navigational intent, enabling the prediction of probabilistic trajectories for planning purposes. In this work, we present a simple yet novel approach for confidence-aware prediction that treats future predictions as particles. This framework is highly parallelized and accelerated on an graphics processing unit (GPU). As a result, this enables longer-term predictions at a frequency of 125 Hz and can be easily extended for multi-human predictions. Compared to existing methods, our implementation supports finer prediction time steps, yielding more granular trajectory forecasts. This enhanced resolution allows motion planners to respond effectively to subtle changes in human behavior. We validate our approach through real-world experiments, demonstrating a robot safely navigating among multiple humans with diverse navigational goals. Our results highlight the methods potential for robust and efficient human-robot coexistence in dynamic environments.
Paper Structure (14 sections, 12 equations, 2 figures, 1 table, 1 algorithm)

This paper contains 14 sections, 12 equations, 2 figures, 1 table, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. PROBLEM FORMULATION
  3. METHODOLOGY
  4. Model-Based Confidence-Aware Human Prediction
  5. Parallelizable Human-Prediction
  6. Safe Trajectory Planning
  7. Global Planner
  8. Local Planner
  9. EXPERIMENTS
  10. Numerical Simulation
  11. Safe Human-Aware Planning in the Real World
  12. ANA$^*$
  13. MPPI
  14. LIMITATIONS & CONCLUSION

Figures (2)

  • Figure 1: In the experiment, the humans are tasked to navigate in an $9.6 \times 5.4$ meter room as demonstrated in the left picture. The picture on the right shows an RVIZ display of the human predictions and their goals (red).
  • Figure 2: Evolution of Human Predictions and visualized MPPI trajectories over three snapshots of the experiment. The coloured visualizations indicate the future predicted occupancies. (a) A human reaches a goal. (b) As the human starts to walk towards another goal, the predictions change to be the mean of the two goals (red). As a result, the robot slightly quickly adjusts it's trajectory. (c) The human decides to walk to the lower left goal. Hence, the robot's plan largely remains the same.