Towards Adaptive Planning of Assistive-care Robot Tasks

TL;DR

This paper presents an adaptive path planning framework for assistive-care robots that combines environment modeling, dynamic path finding, and human movement prediction to enable real-time re-planning in complex environments.

Contribution

It introduces a novel adaptive planning framework integrating graph-based environment modeling, probabilistic human movement prediction, and dynamic re-planning for assistive robots.

Findings

Framework successfully navigates simulated environments

Predictive module accurately estimates human movement

Enables real-time adaptive re-planning in assistive scenarios

Abstract

This 'research preview' paper introduces an adaptive path planning framework for robotic mission execution in assistive-care applications. The framework provides a graph-based environment modelling approach, with dynamic path finding performed using Dijkstra's algorithm. A predictive module that uses probabilistic model checking is applied to estimate the human's movement through the environment, allowing run-time re-planning of the robot's path. We illustrate the use of the framework for a simulated assistive-care case study in which a mobile robot navigates through the environment and monitors an end user with mild physical or cognitive impairments.