Exploring Granger causality in dynamical systems modeling and performance monitoring

Saha, Homagni; Liu, Chao; Jiang, Zhanhong; Sarkar, Soumik

doi:10.1109/cdc.2018.8619530

Cited by 3 publications

(5 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Free energy has been calculated on the entire symbolic time series which includes events before and after the anomaly. It is seen that just based on free energies, STPN and T-STPN frameworks are not able to indicate introduction of anomalies based on jumps in free energy (a fact not investigated in (Saha et al (2018)). However G-STPN shows sudden jumps at 16000-18000 time window for controller hack, and the time windows 12000-14000; 16000-18000 for communication delays.…”

Section: Performance Analysismentioning

confidence: 97%

“…We demonstrate and validate our algorithm in the case of a 7 degree of freedom manipulator using multiple cameras that are able to capture multi-angle views of an object of interest for tracking an object for pick and place task. (Granger (1988)) or for identifying key features in large-scale CPSs for anomaly detection (Saha et al (2018)).…”

Section: Task Driven Information Fusionmentioning

confidence: 99%

“…We use transfer entropy as the metric for causality. However, justifying why transfer entropy is an appropriate replacement for the original formulation of Granger causality is beyond the scope of the paper and readers are referred to (Saha et al (2018)) for further clarification.…”

Section: Formulation Of Granger Causalitymentioning

confidence: 99%

“…2007)), with applications in diagnostics and root-cause analysis of physical faults and cyber anomalies in CPSs( Liu et al (2017b),Saha et al (2018)), residential energy disaggregation(Liu et al (2018a)), building occupancy detection(Tan et al (2019b)) and wind energy prediction(Jiang et al (2017)). While the method has been shown to be effective in practice, so far there has been no rigorous analysis on whether it is able to capture causality among observations from the subsystems.…”

mentioning

confidence: 99%

“…Note that in a preliminary study(Saha et al (2018)), we used the term G-STPN for the STPN variant with transfer entropy. However, in this section we refer to this variant as T-STPN as we call our new framework as Granger STPN (G-STPN) where we propose a new metric comparable to Granger causality.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Information fusion for decision making

Saha

View full text Add to dashboard Cite

Making intelligent decisions in autonomous systems often requires the use of all available information. This information generally comes from multiple sources and may be expressed through multiple modalities. This thesis presents a collection of different approaches for processing and fusing sensor data that have been developed for systems that enjoy multiple sensors and/or sensing modalities. Within these approaches we focus on information fusion in the form of emerged communication for team tasks in multi-agent systems in the first part. After that we investigate information fusion in the context of active sensing by treatment of an agent as a hybrid switched system. Then we look at information fusion in a probabilistic graphical modeling framework. Here information from multiple sensors can be fused through joint probability distributions borrowing notions of causality. These notions of causality are used for deriving intelligent fusion rules that can either help in performance monitoring/anomaly detection or data abstraction by hierarchically combining information across multiple sensors. Finally we propose a novel modular approach for fusion of information from vision (RGB images) and language modalities (user input text) in order for a robot to follow and execute natural language instructions in a household environment. Our proposed technique achieves state of the art accuracy in the popular, realistic Ai2Thor environment based on a Unity based simulator.

show abstract

Section: Performance Analysismentioning

confidence: 97%

Section: Task Driven Information Fusionmentioning

confidence: 99%

Section: Formulation Of Granger Causalitymentioning

confidence: 99%