Using Particle Filter and Machine Learning for Accuracy Estimation of Robot Localization

Eder, Matthias; Reip, Michael; Steinbauer, Gerald

doi:10.1007/978-3-030-22999-3_60

Cited by 2 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 2 summarizes the selection of these features. For a complete description of the list we refer to [36]. During the recording of the training data feature estimations are recorded and labelled with their localization state, as described in (1).…”

Section: Feature Selectionmentioning

confidence: 99%

See 1 more Smart Citation

Creating a robot localization monitor using particle filter and machine learning approaches

Eder

Reip²,

Steinbauer

2021

Appl Intell

Self Cite

View full text Add to dashboard Cite

Robot localization is a fundamental capability of all mobile robots. Because of uncertainties in acting and sensing, and environmental factors such as people flocking around robots, there is always the risk that a robot loses its localization. Very often behaviors of robots rely on a reliable position estimation. Thus, for dependability of robot systems it is of great interest for the system to know the state of its localization component. In this paper we present an approach that allows a robot to asses if the localization is still correct. The approach assumes that the underlying localization approach is based on a particle filter. We use deep learning to identify temporal patterns in the particles in the case of losing/lost localization. These patterns are then combined with weak classifiers from the particle set and sensor perception for boosted learning of a localization estimator. Through the extraction of features generated by neural networks and its usage for training strong classifiers, the robots localization accuracy can be estimated. The approach is evaluated in a simulated transport robot environment where a degraded localization is provoked by disturbances cased by dynamic obstacles. Results show that it is possible to monitor the robots localization accuracy using convolutional as well as recurrent neural networks. The additional boosting using Adaboost also yields an increase in training accuracy. Thus, this paper directly contributes to the verification of localization performance.

show abstract

Section: Feature Selectionmentioning

confidence: 99%

“…Thus, a total of 60.000 iterations would have been sufficient to train this neural network. We refer to [36] for more detailed results.…”

Section: Validation Of the Neural Networkmentioning

confidence: 99%