System interdependence analysis for autonomous mobile robots

Abstract

Autonomous mobile robots are deployed in a variety of application domains, resulting in scenario specific implementations. However these systems share common components responsible for perception, path planning and task execution. In order to find a formal way to identify the influence of the environmental complexity to the used methods, an approach for quantitative system interdependence analysis is introduced. The coherence between several performance indicators of different system components, as well as the influence of environmental parameters on the system, are learned and quantitatively evaluated. Performance evaluation of an autonomous robot navigating in two different urban environments is conducted and presented results demonstrate the applicability of the proposed approach.

A Framework for Information Distribution, Task Execution and Decision Making in Multi-Robot Systems

“…Therefore we utilize the system interdependence analysis described in [25], where the metric interdependencies are modeled probabilistically by searching and training a Bayesian Network structure. The Bayesian Network enables the quantification of the mutual metric interdependencies whereby the relation between task-specific and scenario-specific metrics is obtained.…”

Interconnected performance optimization in complex robotic systems

“…In this respect, we utilize the probabilistic system interdependence analysis described in [4] to learn a probabilistic model of the metric interdependencies. First a set of performance metrics needs to be specified for each task the robot can perform.…”

“…The metric values are first discretized and then a Bayesian Network (BN) structure, which best reflects the interdependencies between the metrics, is searched. As proposed in [4], a combination of a Markov Chain Monte Carlo [12] and K2 [13] search is used to identify the best structure. As quality measure the Bayesian Information Criterion (BIC) [14] is used.…”

“…The performance of a localization subtask for example is commonly measured by the uncertainty of its pose estimate [3]. Especially in mobile robots the localization performance indirectly influences the global metric [4], such as completion time or the number of scored goals. Another case are subtasks with conflicting goals.…”

“…First, a probabilistic model of the interdependencies is learned. The generation of the model is based on [4] which provides a quantitative determination of the interdependencies within a set of arbitrary metrics through the usage of Bayesian Networks (BN). Furthermore, a method to tightly connect the planning with the task execution of a robotic system is proposed.…”