how to build an autonomous anything - mathworks · predicting customer behavior using big data...
Post on 09-Jun-2020
5 Views
Preview:
TRANSCRIPT
1© 2015 The MathWorks, Inc.
How to build an autonomous anything
Michelle HirschHead of MATLAB Product ManagementMathWorks
2
3
4
5
6
7
8
Autonomous Technology
9
Autonomous
Having the power for self-governance
10
Autonomous Technology
Provides the ability of a system to act independently of direct human controlunder unrehearsed conditions
11
12
Capabilities of an Autonomous System
Sense
13
Capabilities of an Autonomous System
Sense
Perceive
14
Capabilities of an Autonomous System
Learning AlgorithmsOptimization
Sense
Perceive
Decide & Plan
15
Capabilities of an Autonomous System
Control Algorithms
Sense
Perceive
Decide & Plan
Act
16
Autonomous Technology Transfers Responsibility to Computers
Human
Computer
Res
pons
ibili
ty
Degree of Autonomy
17
Cost of rig: >$1MRepair cost: $100,000
Cost of valve: $200
18
19
20
21
Autonomous Service for Predictive Maintenance
Which sensor values should they use?
Decide & Plan
Act
Perceive
Sense
Pressure
Other variables
Vibration Timing
Temperature
22Normal Operation Maintenance NeededMonitor Closely
Autonomous Service for Predictive Maintenance
Sense
Act
Perceive
Decide & Plan
23
Machine Learning or Deep Learning?
1. Normal2. Monitor3. Maintain
Classification OutputFeature Extraction
Correlation Analysis
Feature Extraction & Classification
Sensor 1Sensor 2…Sensor 25
Sensor aSensor bSensor c
Sensor 1Sensor 2…Sensor 25
Machine Learning Approach
Deep Learning Approach
1. Normal2. Monitor3. Maintain
Output
24
What are the best predictors?
Data-driven
25
What are the best predictors?
Data-driven Model-driven
Jet Engine Monitoring
26
What are the best predictors?
Data-driven Model-driven
Jet Engine Monitoring Find out more about predictive maintenance:Build predictive maintenance algorithms using physical models
Demo Station
Find out more about machine learning:Machine Learning Simplified
Paola JaramilloTrack B 13:15 – 13:45
27
Bazille’s StudioFrederic Bazille (Paris, 1870)
Shuffleton’s BarbershopNorman Rockwell (Vermont, 1950)
28
Bazille’s StudioFrederic Bazille (Paris, 1870)
Shuffleton’s BarbershopNorman Rockwell (Vermont, 1950)
29
Autonomous Artistic Style ClassificationRutgers University
Image Feature
ExtractionVisual Features
Genre Classifier
(SVM)
Artist Classifier
(SVM)
Style Classifier
(SVM)
Style:Regionalism
Genre:Interior
Artist:Rockwell
Machine Learning
Classification
Sense
Perceive
Decide & Plan
Act
30
Where to add autonomy with perception?
Analyze more data Reduce bias Reduce variability Save time Improve performance
Determine Loudspeaker
Quality
Virtual Semiconductor Manufacturing Calibration
31
Autonomous Glucose Level Management
32
Autonomous Glucose Level ManagementBigfoot Biomedical
Sense
Perceive
Decide & Plan
ActContinuous
Glucose Monitor
Target Glucose
Level
+ -
Insulin PumpPerson
33
Autonomous Glucose Level ManagementBigfoot Biomedical
Continuous Glucose Monitor
Target Glucose
Level
+ -
Insulin PumpPerson
+ +
Mobile App
Sense
Act
Perceive
Decide & Plan
34
Autonomous Glucose Level ManagementBigfoot Biomedical
Continuous Glucose Monitor
Insulin Pump
Sense
Act
Decide & Plan
Perceive
Person
Simulink, Stateflow,Polyspace
Mobile App
Target Glucose
Level
+ -
+ +
35
Autonomous Glucose Level ManagementBigfoot Biomedical
Continuous Glucose Monitor
Target Glucose
Level
+ -
Insulin Pump
+ +
Person
Perceive
Decide & Plan
Act
Sense
Mobile App
36
Autonomous Glucose Level ManagementBigfoot Biomedical
Person
Perceive
Decide & Plan
Act
Sense
Continuous Glucose Monitor
Target Glucose
Level
+ -
Insulin Pump
+ +
Mobile App
20171 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
20181 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
37
Virtual ClinicGenerating data through simulation
38
Virtual ClinicScaling computations to simulate 50 million patients a day
39
Where will you get your data?
Simulation Public repositories In the lab In the field Internet of Things (IoT)
40
Where will you get your data?
Simulation Public repositories In the lab In the field Internet of Things (IoT)
Find out more:Predicting Customer Behavior Using Big Data Analytics with MATLAB in the CloudRachid el Mimouni, NLETrack B 15:15 – 15:45
Find out more:Verification Techniques for Model and Code
Paul LambrechtsTrack A 15:15 – 15:45
41
CNH Develops Intelligent Filling System for Forage Harvesters
42
Autonomous Trailer Filling
Control Algorithms
Perceive
Act
Sense
Decide & Plan
43
Autonomous Trailer Filling
Control AlgorithmsAct
Decide & Plan
Perceive
Sense
44
45
Autonomous Trailer Filling
Computer vision and controls algorithms
Act
Perceive
Decide & Plan
Sense 3D Cameras
CAN
ECU
Actuators
46
• Driver Input• Visualization
Autonomous Trailer Filling
Act
Perceive
Decide & Plan
Sense
Computer vision and controls algorithms
Vehicle Display Controller
ECU
Actuators
3D Cameras
CAN
47
• Driver Input• Visualization• Computer
Vision• Controls
Autonomous Trailer Filling
Act
Perceive
Decide & Plan
Sense
Computer vision and controls algorithms
Vehicle Display Controller
ECU
ActuatorsEmbedded Coder
3D Cameras
CAN
48
Autonomous Trailer Filling
Monitoring
• Driver Input• Visualization• Computer
Vision• Controls
Act
Perceive
Decide & Plan
Sense 3D Cameras
ECU
Actuators
Vehicle Display Controller
CAN
49
How will you put it into production?
System Architecture
Embedded systems Enterprise systems HMIs
50
How will you put it into production?
System Architecture
Embedded systems Enterprise systems HMIs
Find out more:MATLAB Analytics in Enterprise Applications
Ionut BarbuTrack B 14:45 – 15:15
51
52
How to build an autonomous anything
Focus on Perception
• Data-driven• Model-driven
• Reduce to actionable data• Take advantage of Big Data• Use simulation to supplement available data
• Address the architecture• Leverage Model-Based Design for embedded• Automate integration with enterprise IT systems
• Look for autonomy in creative places • Do more than manually possible
Use the Best Predictors
Get the Right Data
Go to Production
53
What is yourautonomous anything?
top related