19
Webinar: Automating Complex Airport Operations with WSO2 Middleware Platform Miyuru Dayarathna and Ramindu De Silva

Automating Complex Airport Operations with WSO2 Middleware Platform

Embed Size (px)

Citation preview

Page 1: Automating Complex Airport Operations with WSO2 Middleware Platform

Webinar: Automating Complex Airport Operations with WSO2 Middleware Platform

Miyuru Dayarathna and Ramindu De Silva

Page 2: Automating Complex Airport Operations with WSO2 Middleware Platform

Introduction● Operations conducted in a typical airport are

diversified and often span multiple information systems.

● Examples include,○ flight check-in○ flight information maintenance○ cleaning crew operations○ traffic control (air/ground) ○ etc.

2

Page 3: Automating Complex Airport Operations with WSO2 Middleware Platform

Problem● Information systems involved in airport operations are

mostly isolated [1].● We need comprehensive

enterprise middleware platforms over isolated solutions.● Why?

○ Reduced operation cost○ Improved service quality○ Time-based

competitiveness○ New products and services

3

[1] Norman J. Ashford, Saleh Mumayiz, Paul H. Wright (2011), Airport Engineering: Planning, Design and Development of 21st Century Airports, John Wiley & Sons, 2011

Page 4: Automating Complex Airport Operations with WSO2 Middleware Platform

WSO2 Data Analytics Server

4

Page 5: Automating Complex Airport Operations with WSO2 Middleware Platform

Real-time Analytics : WSO2 Complex Event Processor

5

Page 6: Automating Complex Airport Operations with WSO2 Middleware Platform

Predictive Analytics :WSO2 Machine Learner

6

Page 7: Automating Complex Airport Operations with WSO2 Middleware Platform

Solution

Use a comprehensive middleware platform to develop a unified solution.o Front-end services

o Provide real-time information on average wait time

o Back-end serviceso Material serviceso Fleet services

7

Page 8: Automating Complex Airport Operations with WSO2 Middleware Platform

Front-end information services : Providing real-time information• Information such as wait time are

important for individual travellers.

8

Page 9: Automating Complex Airport Operations with WSO2 Middleware Platform

Providing real-time information to individuals● We have implemented an iBeacon based wait

time prediction application.

9

WSO2 DAS

Reorder Kalman Filter

Average beacon locations

Update waiting time of specific areas

Publisher

Receiver

GeoDashboard

Supervisor

Trajectory Smoothing + Waiting time calculation

Page 10: Automating Complex Airport Operations with WSO2 Middleware Platform

Re-ordering

● Out-of-order events are possible in most of the event processing scenarios.

● Multiple approaches to deal with the disorder introduced by the out-of-order events exist.○ Buffer-based techniques○ Punctuation-based techniques○ Speculation-based techniques○ Approximation-based techniques

10

Page 11: Automating Complex Airport Operations with WSO2 Middleware Platform

Re-ordering : k-slack

● Uses a buffer to sort tuples from the input stream in ascending timestamp order before presenting them to the query operator.

● k-slack uses a buffer of length K to delay an event for at most K time units

11

timestamp > greatestTimestamp

timeDifference > k

timeDifference < MAX_K

timeDifference = greatestTimestamp - minTimestamp

k = timeDifference k = MAX_K

entry.getKey() + k <= greatestTimestamp

Take next element

Emit element

Has more elements?yes

yes

yes

yes

yes

Page 12: Automating Complex Airport Operations with WSO2 Middleware Platform

Kalman Filter

● Estimate the real values from the measured values in-order to smooth the trajectory.

● Given position and velocity sensor readings (Zk), we need to estimate the real reading (X’^k).

● Kalman filter is a set of mathematical equations.

12

X’^k

Zk

Page 13: Automating Complex Airport Operations with WSO2 Middleware Platform

Providing real-time information to individuals - Contd.

13

Passenger trajectory before Kalman filtering Passenger trajectory after Kalman filtering

Page 14: Automating Complex Airport Operations with WSO2 Middleware Platform

Kalman Filter (Contd.)• Calculate the best estimate (i.e.,

probably the real reading) X’^k and its covariance matrix P’k

• Updated Kalman gain matrix (K’k) need to be calculated using the covariance matrices of the previous estimates

• X’^k and P’k values are fed back to the Kalman filter in the next round of predict or update as many times as required.

X’^k = X^k-1 + Kk-1 ( Zk - H*X^k-1) ---------------------------------- (1)

PredictionX^k = A*X^k-1 ----------------------------------------------------------- (2) Pk = A*Pk-1AT ---------------------------------------------------------- (3)UpdateS = H*Pk*HT+R ------------------------------------------------------- (4) K’k = Pk*HT*S-1 ------------------------------------------------------- (5) X’^k = X^k + K’k ( Zk - H*X^k) ------------------------------------ (6) P’k = Pk - K’k*H*Pk ------------------------------------------------- (7)

Where,A = [1 timeDifference; 0 1]

Initially, the time difference is assumed as 0X = [previouslyEstimatedValue; ChangingRate]

Initially, the previously estimated value was assumed as the initial measured valueP = [1000 0; 0 1000]H = [1 0; 0 1]R = [standardDeviationOfNoise 0; 0 standardDeviationOfNoise]

standardDeviationOfNoise = 0.01

14

Page 15: Automating Complex Airport Operations with WSO2 Middleware Platform

Front-end Information Services :Predicting the service time

15

WSO2 DASCustomer

Persisted Wait-time

stream

DASTable Access

Linear RegressionAnalysis

Model REST API

Wait time information stream

WSO2 ML

Mobile App

Waiting time prediction

Page 16: Automating Complex Airport Operations with WSO2 Middleware Platform

Back-end Information Services : Integrating diversified information - Airplane Maintenance

16

Page 17: Automating Complex Airport Operations with WSO2 Middleware Platform

Back-end Information Services : Integrating diversified information - Maintenance crew management

• Technicians - need to quickly access timely information pertaining to the process of aircraft maintenance.

• Cleaning crew - need to indicate the start/end times of the cleaning process so that the airline officers can start boarding process.

17

Page 18: Automating Complex Airport Operations with WSO2 Middleware Platform

Conclusion

• Typical airport operations are diversified and often require interaction of multiple disconnected information systems.

• We described how WSO2’s comprehensive middleware platform could be leveraged to create integrated, seamless solution for airport operations.

• DAS’s batch and interactive analytics could also be utilized in this process in future.

18