Probabilistic Hybrid Action Models for Predicting Concurrent Percept-driven Robot Behavior

TL;DR

This paper introduces Probabilistic Hybrid Action Models (PHAMs), a new approach for predicting complex, concurrent, percept-driven robot behaviors with probabilistic accuracy, enabling better planning and control of autonomous robots.

Contribution

The paper presents PHAMs, a formal probabilistic model for concurrent robot behavior prediction, along with an efficient inference method and validation through robotic planning experiments.

Findings

PHAMs generate probably, qualitatively accurate predictions.

The inference method is resource-efficient and effective.

Application to robot office courier planning demonstrates practical utility.

Abstract

This article develops Probabilistic Hybrid Action Models (PHAMs), a realistic causal model for predicting the behavior generated by modern percept-driven robot plans. PHAMs represent aspects of robot behavior that cannot be represented by most action models used in AI planning: the temporal structure of continuous control processes, their non-deterministic effects, several modes of their interferences, and the achievement of triggering conditions in closed-loop robot plans. The main contributions of this article are: (1) PHAMs, a model of concurrent percept-driven behavior, its formalization, and proofs that the model generates probably, qualitatively accurate predictions; and (2) a resource-efficient inference method for PHAMs based on sampling projections from probabilistic action models and state descriptions. We show how PHAMs can be applied to planning the course of action of an autonomous robot office courier based on analytical and experimental results.