predictive analytics in sensor (iot) data · predictive analytics in sensor (iot) data andrás...
TRANSCRIPT
![Page 1: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/1.jpg)
Predictive analytics in Sensor (IoT) Data
András Benczúr, head, Informatics Lab
Anna Mándli, Ph.D. student, Bosch
Róbert Pálovics, postdoctoral researcher, recommenders, online machine learning
Bálint Daróczy, postdoctoral researcher, computer vision, deep learning
March 20, 2018
![Page 2: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/2.jpg)
Big Data – Lendület kutatócsoport
• Összesen 6 műszaki/informatika Lendület csoport
• Lendület támogatás a költségvetés ~15%-a, ~ 40% közvetlen szerződés, 20-20% hazai és EU, 5% SZTAKI belső erőforrás
• Teljes innovációs lánc, kutatástól alkalmazásokig
2018. 01. 09. Informatika Kutató Laboratórium 2
![Page 3: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/3.jpg)
Alkalmazási területeink
3 July 7, 2017 Manufacturing IoT Analytics
• AEGON
– 40 millió ügyfélrekord duplikátum-mentesítése az éles ügyfél rendszerben
– Csalásfelderítő eszköz
• OTP
– Gépi tanulási eljárások alkalmazásában elnyert tender
– Orosz, Román ügyfelekre hitelnemfizetés, hitelkártya viselkedés modellezés
– 9Md tranzakció hálózatában kockázat-terjedés modellezés
• Ericsson
– Mobil session drop, Quality of Experience predikció
– Data Streaming analitika, világméretű IoT Data Hub prototípus
• Bosch
– Gyártósori szenzor adatok alapján minőségi problémák előrejelzése, root cause analízis
• Telekom, Vodafone, Clickshop: kereső technológia
![Page 4: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/4.jpg)
Use case 1: scrap rate prediction in transfer molding
![Page 5: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/5.jpg)
Cleaning cycles
5 July 7, 2017 Manufacturing IoT Analytics
![Page 6: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/6.jpg)
Scrap rate prediction task
6 July 7, 2017 Manufacturing IoT Analytics
![Page 7: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/7.jpg)
Transfer Pressure
Vacuum Pressure
• Multiple time series, 500+ measurements in each: many, many data points
• Features of mean, variance, differentials, distribution of transfer graph data points
• In the end, ~50 features in the final classification data
Data
July 7, 2017 Manufacturing IoT Analytics 7
Filling pressure
![Page 8: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/8.jpg)
Time series of time series
8 July 7, 2017 Manufacturing IoT Analytics
![Page 9: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/9.jpg)
Series of pressure series
• Filling pressure measurements
• Red to yellow color scale: measurement 1 – 50
• Vertical lines: new charges
• Important to notice
– Values are lower after machine cleaning, and later they typically increase
– Pressure is indirectly measured on the plunger
– Contamination modifies pressure measurement
9 July 7, 2017 Manufacturing IoT Analytics
![Page 10: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/10.jpg)
Summary
Machine logs
• BottomPreheatTemperature1-6
• BottomToolTemperature
• UpperToolTemperature
• LoaderTemperature
• PreheatTime
• toolData[1-2].value
• TransferPressGraphPos1-150
• TransferSpeed1-10
• TransferTime
• TransferVacuum
Facts:
6-10,000 products/hour
Few 100 data points per product
100> failed products in a day
Available for several months:
Delamination, Void, … failures
• First line reject
• Second line reject
• Failures at later stages
Root Cause Analysis
Transfer pressure measured indirectly
• includes friction from valves
• certain points in the time series indicate
contamination
Vacuum pressure has no direct effect, but …
• Differences in trays: calibration problems
• Vacuum drop speed: leakage and blinding
• May result in less effective cleaning?
Result in variables that affect the production
in indirect way that needs to be understood
July 7, 2017 Manufacturing IoT Analytics 10
![Page 11: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/11.jpg)
Use Case 2: Mobile radio session drop
![Page 12: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/12.jpg)
Data is based on eNodeB CELLTRACE logs from a live LTE (4G) network
RRC connection setup /
Successful handover into the cell
UE context release/
Successful handover out of the cell
Per UE measurement reports
(RSRP, neighbor cell RSRP list)
Configurable period
Not available in this analysis
Per UE traffic report
(traffic volumes, protocol events (HARQ, RLC))
Per radio UE measurement
(CQI, SINR)
Period: 1.28s
START END
![Page 13: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/13.jpg)
Root-cause analysis: Cause of Release and Drop
0.6% of sessions are dropped
![Page 14: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/14.jpg)
Latency critical communication with unreliable links
Detect network issues, predict problems ahead in time to prepare alternate control route
Make stable and reliable re-routing decisions in case of jitter
Reroute traffic for robot control and other real time applications
![Page 15: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/15.jpg)
Examples for low quality and loss
High quality Low quality
Dropped Normal
Time (ms) Time (ms)
dB
m
dB
m
![Page 16: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/16.jpg)
Use case 2 B: Session (call) drop by Smartphone logs
• Csaba Sidló • Mátyás Susits • Barnabás Balázs
![Page 17: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/17.jpg)
Accelerometer, 5 different devices
Acc X
A
cc Y
A
cc Z
Time (ms)
Logging app and time series sample for illustration
![Page 18: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/18.jpg)
Heating until drop
(by CPU load + lamp)
Drop at 70C
Overheated phone test until turn off
![Page 19: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/19.jpg)
Normal calls
still up to 65C
Overheated phone test until turn off
감사합니다!
Thank You!
!תודה
![Page 20: Predictive analytics in Sensor (IoT) Data · Predictive analytics in Sensor (IoT) Data András Benczúr, head, Informatics Lab Anna Mándli, Ph.D. student, Bosch Róbert Pálovics,](https://reader036.vdocuments.us/reader036/viewer/2022081614/5fbfe73cafe12e0fc9728baa/html5/thumbnails/20.jpg)
Kérdések?