Automated Robot Behavior Recognition

Abstract

Automated recognition of the behavior of robots is increasingly needed in a variety of tasks, as we develop more autonomous robots and general information processing agents. For example, in environments with multiple autonomous robots, a robot may need to make decisions based on the behavior of the other robots. As another interesting example, an intelligent narrator agent observing a robot will need to automatically identify the robot's behaviors. In this paper, we introduce a novel framework for using Hidden Markov Models (HMMs) to represent and recognize strategic behaviors of robotic agents. We first introduce and characterize the perceived signal in terms of behavioral-relevant state features. We then show how several HMMs capture different defined robot behaviors. Finally we present the HMMbased recognition algorithm which orchestrates and selects the appropriate HMMs in real time. We use the multi-robot robotic soccer domain as the substrate of our empirical validation, both in simulation and using real robots. Robots will then adapt their behaviors as a function of the autonomously recognized behavior of the other agents, either teammates or opponents.

A Formalization of the Coach Problem

“…The latter approach is sometimes called 'opponent modelling' and used to learn an optimal strategy against the modelled type of agent [1,8,38,39]. Although plan recognition was first introduced in the field of logical inference by [25], the current state of the art is mainly based on probabilistic methods like Bayesian Networks [6,7,9] and Markov Models [6,4,21,10,46,26]. For example in [39], simulated robot soccer players adapt their positions strategically in adaptive response to the opponent's behaviour throughout the game.…”

Creating User Profiles from a Command-Line Interface: A Statistical Approach

“…Therefore, HMMs are also used in the environment we propose in this research. HMMs have been used for recognizing automated robot behaviors [20] and recently, for behavior understanding from video streams [21]. In addition, Lane [22] demonstrates the use of HMMs for user profiling in the domain of anomaly detection using a data set very similar to the set used in our research.…”

Robust Real-Time Human Activity Recognition from Tracked Face Displacements

“…Recognizing the behaviors of individual robotic (non-human) agents has been studied in [8]. Robots playing soccer against other robots [9] would greatly benefit by being able to classify the different behavioral patterns observed by their opponents.…”