Sensor Failure Detection, Identification, and Accommodation Using Fully Connected Cascade Neural Network

Hussain, Saed; Mokhtar, Maizura; Howe, Joe

doi:10.1109/tie.2014.2361600

Cited by 92 publications

(52 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Accordingly, we applied the cascade neural network modeling approach to predict the flight departure delay first and then use it as the input to predict the flight arrival delay, as shown in Fig. . Set of inputs are shown as in Table .…”

Section: Flight Delay Forecasting Method—cascade Neural Networkmentioning

confidence: 99%

Cascading Delay Risk of Airline Workforce Deployments with Crew Pairing and Schedule Optimization

Chung

Chan

2016

Risk Analysis

View full text Add to dashboard Cite

This article concerns the assignment of buffer time between two connected flights and the number of reserve crews in crew pairing to mitigate flight disruption due to flight arrival delay. Insufficient crew members for a flight will lead to flight disruptions such as delays or cancellations. In reality, most of these disruption cases are due to arrival delays of the previous flights. To tackle this problem, many research studies have examined the assignment method based on the historical flight arrival delay data of the concerned flights. However, flight arrival delays can be triggered by numerous factors. Accordingly, this article proposes a new forecasting approach using a cascade neural network, which considers a massive amount of historical flight arrival and departure data. The approach also incorporates learning ability so that unknown relationships behind the data can be revealed. Based on the expected flight arrival delay, the buffer time can be determined and a new dynamic reserve crew strategy can then be used to determine the required number of reserve crews. Numerical experiments are carried out based on one year of flight data obtained from 112 airports around the world. The results demonstrate that by predicting the flight departure delay as the input for the prediction of the flight arrival delay, the prediction accuracy can be increased. Moreover, by using the new dynamic reserve crew strategy, the total crew cost can be reduced. This significantly benefits airlines in flight schedule stability and cost saving in the current big data era.

show abstract

Section: Flight Delay Forecasting Method—cascade Neural Networkmentioning

confidence: 99%

Cascading Delay Risk of Airline Workforce Deployments with Crew Pairing and Schedule Optimization

Chung

Chan

2016

Risk Analysis

View full text Add to dashboard Cite

show abstract

“…In order to detect sensor failure, we will be measuring uncertainty in NN models, in comparison to the more traditional approach of sensor fault detection before passing sensor information to a NN. There exists an extensive literature on sensor failure detection [12], [13], [14] that demonstrates its application in various fields. However, this approach requires knowledge of the expected sensor outputs to determine whether a reading is normal or faulty.…”

Section: Introductionmentioning

confidence: 99%

Ensemble Bayesian Decision Making with Redundant Deep Perceptual Control Policies

Lee

Wang

Vlahov

et al. 2019

2019 18th IEEE International Conference on Machine Learning and Applications (ICMLA)

View full text Add to dashboard Cite

This work presents a novel ensemble of Bayesian Neural Networks (BNNs) for control of safety-critical systems. Decision making for safety-critical systems is challenging due to performance requirements with significant consequences in the event of failure. In practice, failure of such systems can be avoided by introducing redundancies of control. Neural Networks (NNs) are generally not used for safety-critical systems as they can behave in unexpected ways in response to novel inputs. In addition, there may not be any indication as to when they will fail. BNNs have been recognized for their ability to produce not only viable outputs but also provide a measure of uncertainty in these outputs. This work combines the knowledge of prediction uncertainty obtained from BNNs and ensemble control for a redundant control methodology. Our technique is applied to an agile autonomous driving task. Multiple BNNs are trained to control a vehicle in an end-to-end fashion on different sensor inputs provided by the system. We show that an individual network is successful in maneuvering around the track but crashes in the presence of unforeseen input noise. Our proposed ensemble of BNNs shows successful task performance even in the event of multiple sensor failures.

show abstract

“…On the other hand, active fault-tolerant control has attracted the attention of researchers over the past decades, [4][5][6][7][8][9][10][11] which has the capability of reacting to the uncertain system faults actively. By collecting and comparing the measurements from both faulty sensors and correct sensors, the failed sensors could be detected and be recovered, [16][17][18][19] if the control system has multiple sensors to measure the same signal or different signals with known inner connections. 12 Recently, a direct adaptive active fault-tolerant control approach has been developed, which does not use explicit fault detection and isolation but directly adjust the fault compensation controller parameters for control reconfiguration.…”

Section: Introductionmentioning

confidence: 99%

“…15 For systems with sensor failures, one branch of approaches is based on neural network. By collecting and comparing the measurements from both faulty sensors and correct sensors, the failed sensors could be detected and be recovered, [16][17][18][19] if the control system has multiple sensors to measure the same signal or different signals with known inner connections. Recently, some new approaches are proposed to avoid the sensor redundancy requirement, which may confine the application of the aforementioned approaches.…”

Section: Introductionmentioning

confidence: 99%

Adaptive state‐feedback control with sensor failure compensation for asymptotic output tracking

Song

Tao

2018

Adaptive Control & Signal

View full text Add to dashboard Cite

This paper addresses a new adaptive output tracking problem in the presence of uncertain plant dynamics and uncertain sensor failures. A new unified nominal state-feedback control law is developed to deal with various sensor failures, which is directly constructed by state sensor outputs. Such a new state-feedback compensation control law is able to ensure the desired plant-model matching properties under different failure patterns. Based on the nominal compensation control design, a new adaptive compensation control scheme is proposed, which guarantees closed-loop signal boundedness and asymptotic output tracking. The new adaptive compensation scheme not only expands the sensor failures types that the system could tolerate but also avoids some signal processing procedures that the traditional fault-tolerant control techniques are forced to encounter. A complete stability analysis and a representative simulation study are conducted to evaluate the effectiveness of the proposed adaptive compensation control scheme.KEYWORDS model reference adaptive control, plant-model output matching, sensor failure compensation, state feedback OverviewEfforts have been devoted to handle the undesired effects on system performance brought by those adverse conditions. In this section, we will first give a brief review of fault-tolerant control, especially a review of techniques dealing with sensor faults and then discuss the research motivation of this paper. 130

show abstract

Sensor Failure Detection, Identification, and Accommodation Using Fully Connected Cascade Neural Network

Cited by 92 publications

References 44 publications

Cascading Delay Risk of Airline Workforce Deployments with Crew Pairing and Schedule Optimization

Cascading Delay Risk of Airline Workforce Deployments with Crew Pairing and Schedule Optimization

Ensemble Bayesian Decision Making with Redundant Deep Perceptual Control Policies

Adaptive state‐feedback control with sensor failure compensation for asymptotic output tracking

Contact Info

Product

Resources

About