Advanced Personal Robot Interaction Research Dr. Jiqiang Song Intel Labs China
[email protected]
1
Heading from CT to RT era
RT = CT + Robot Technology
Computer Technology
Sensing
Perception
AI
Cognition
ME
Action
user
Artificial Intelligence
Mechanical Electronics
2
Human-robot Interaction (HRI)
What is an ideal personal robot? • Siri*? • NAO*? • Atlas*? • Pepper*? • Prof. Ishiguro’s robot clone? • DLR* Rollin’ Justin*? Image source:
3
https://www.aldebaran.com/en
http://www.eng.osaka-u.ac.jp/en/index.html
http://www.bostondynamics.com/
http://www.dlr.de/rmc/rm/en/desktopdefault.aspx/tabid-5471/
We count on robots to solve BIG social problems China demographics at the year 2010
HSR
As of today, 212M Age > 60 Keep increasing 15 years to the peak: 350M
Image source: https://en.wikipedia.org/wiki/ Demographics_of_China 4
Three essentials for personal robot booming Useful
Affordable
¥ Specific functions and usages
5
Reliable
$
Reasonable price & business model
Safe, durable & privacy preserving
Three stages of personal robots We are here!
Connected • • • •
6
Wireless com Home gateway Information delivery Social service
Smart • • • •
Listen & speak See & recognize Telepresence Personalization
Autonomous • • • •
Behavior understanding Emotion understanding Reasoning & planning Reliable & predictable
From Smart to Autonomous: four key pieces Sensor fusion
Smart computing acceleration
7
Distributed heterogeneous computing
Safe & Reliable
#1: Sense beyond human
8
Sensor fusion
Distributed heterogeneous computing
Smart computing acceleration
Safe & Reliable
Intel® RealSenseTM Product (2016)
9
3D Sensing with Intel® RealSenseTM Technology
10
Limitation of Single Sensor
11
Out of range Reflective Dark Material
Image source: https://homes.cs.washington.edu/~xren/publication/3d-mapping-iser-10-final.pdf
Need more sensor input to improve navigation Robot Navigation LiDAR
12
U/S range sensor
IR range sensor
IMU
Odometer
SLAM Fusion for Robot Navigation • Deliver flexible & affordable autonomous navigation solutions in complex home environment
LiDAR UWB VSLAM IMU/ Odom
Sensing 13
SLAM Fusion (Optimized Algorithm)
Localization and Mapping
ID/MD/HW integration 19
Planning and Exploration
Need more sensor input to improve HRI Robot Navigation LiDAR
U/S range sensor
IR range sensor
IMU
Odometer
Robot Interaction RGB camera
14
Depth camera
Microphone array
mmWave scanner
nSense fusion for human/object detection
RGB camera
15
Depth camera
mm-Wave scanner (30GHz)
#2: Compute in the right way
16
Sensor fusion
Distributed heterogeneous computing
Smart computing acceleration
Safe & Reliable
Distributed heterogeneous computing Low-power AOAC system (Always On Always Connected)
High performance smart computing system (CPU+GPU+FPGA)
Real-time motion control system (MCU) 17
Voice and wireless wakeup
AOAC
Voice command
18
Anomaly detection
High-performance smart computing system SIP
Configurable I/O
QPI/UPI
CPU
PCIE Prog I/F
FPGA AFU
Peripherals
PCIE
GPU
19
QPI: Quick Path Interconnect UPI: Ultra Path Interconnect Prog I/F: FPGA programming interface AFU: Accelerator Function Unit
FPGA acceleration for smart computing Applications
• FPGA - Sensing and acting
CNN SIFT
- Cognitive computation
• CPU - Planning and decision - Communication and service
DNN DTW
FFT
Face/Emotion Motion control SSD DDR
FPGA Fuse/Accelerate
Planning
Semantic scene
DDR
CPU PCIE
App/Service
Sensors
Actions 20
ORB
SURF
De-noise
LiDAR mm-Wave IMU MIC Array RealSense camera
Visual-SLAM
Emergency Hands Arms Wheels
20
Wi-Fi BT 5G
#3: Add more intelligence, and FAST! Sensor fusion
Smart computing acceleration
21
Distributed heterogeneous computing
Safe & Reliable
Computer Vision – What is it? Computer Vision (CV) is a field that includes methods for acquiring, processing, analyzing and understanding images and, in general, high-dimensional data from the real world in order to produce numerical or symbolic information Wikipedia.org
Imaging Acquire
Process
Pixel processing
Visual Understanding Analyze
Understand
Object processing
Sample Capabilities
22
Face Detection & Recognition
Emotion Recognition
Text Recognition in the wild
Object Classification/Recognition
Activity Classification/Recognition
Scene Classification/Understanding
Video Classification/Summarization
Classification: Person, Camera Detection Person Camera 22
Action: Taking pictures
Target = human: advanced facial technology
GenX Acceleration
Face detection, morphing
Helen
Happy Female Adult
Emotion recognition
SDK
Intel Leading Face Analysis Technology
23
Dynamic facial expression recognition Recognize 7 basic facial expressions: happy, surprise, fear, disgust, anger, sadness and neutral in real time (100 fps on Intel® CORE CPU)
24
Emotion Recognition in the Wild Challenge Intel won EmotiW2015 in the audio-video based task, and competitors included 74 teams (CMU, UIUC, MSR, etc.) across the world • Task 1: EmotiW 2015 AFEW dataset (train/validation/test) 723/383/539 movie clips with audio, 7 basic facial expressions completely shown in the wild • Task 2: EmotiW 2015 SFEW dataset (train/validation/test) 958/436/372 static images, 7 basic facial expressions completely shown in the wild
Overall Recognition Rate (%) on EmotiW2015 Test Set
25
Methods
AFEW
SFEW
Baseline
39.33
39.13
Winner 2014
50.37
N/A
Intel audiovisual/visual solution
53.80
55.38
Active Vision 3D Slam and Navigation • Offline mapping, Robust 2D/3D combined algorithm • Real-time navigation, active vision
Fast Profile register
Robot Motion
Facing people • Novel robot capability Robot Motion
Social & Environmental Perception
Robust
People detection
Potential target
Confirmation
26
Finding people • Robust 360 degree pose human detection (on sofa, close to wall) • Facing people + Face/Person features
Robot Motion
Following people • Long-term tracking • Robust - deal with distraction from other people
Target = object: from 2D to 3D visual recognition
RGB + Depth
3D Point Cloud (3D-PC)
Depth and 3D information can suppress illumination, occlusion & clutter difficulties in 2D… 27
3D Object Detection Unified framework for 3D Pose Estimation and Object Detection using deep learning • Expect near real-time processing speed on Intel® CORE CPU.
RGB-D pair
Region Proposal
Instance Segmentation
Proposal & Mask
Normal Extraction
Pose CNN
Image Cropping
Object Class & Matched Coarse Pose
Model Model3D-PC 3D-PCin inwhite green Fine Pose & Object Class
Proposal & Mask as Input
Iterative Closest Point (ICP)
Cropped 3-channel normal
Coarse Pose & Model as Initialization
Captured/ICP aligned 3D-PC in white/red Captured/ICP aligned 3D-PC in green/red Bed room picture source: NYU Depth Dataset V2, Indoor Segmentation and Support Inference from RGBD Images, ECCV 2012 28
Robot’s memory: go beyond deep-learning Environment Video, image
Conversation
29
Activity Health
Small data and no labels – hard for DL Day
Week
*,*,*,
*,*,
*,*,*,
Medication
**
*
**
Eating
*
**
*
*
*,*,
*,
*,*,*,
*,*,
Location
Drinking
Month
Year
*,
Health data
*,
*,
Audio event
*,
*,
Environment
**
*
**
*
Conversation
…..
..
…
.
…
Time
Goal: Discover patterns in behavior and analyze association, for behavior prediction, anomaly detection 30
AI + HRI will break through
Human-robot Interaction
Perception
CT + AI
Cognition Low confidence?
Action
31
user
Ambiguity? Ask right questions
#4: No one is hurt
32
Sensor fusion
Distributed heterogeneous computing
Smart computing acceleration
Safe & Reliable
Dangerous: mobile robot without security
33
Safe and reliable
Container SE Linux
Data security
Intel®SGX
Physical security
Intel®TXT
34
Privacy leak
Denial of service
Full-stack security mechanism Container 1
Container 2 Secret
Secret
App: ROS
Apps
Apps Safety Framework
OS: Linux
Container N
Safety Framework
System libs Linux kernel
Secret
…
Apps Safety Framework
SE Linux
Preserve privacy
Protect apps’ codes and data Access control
MAC Policy
System integrity HW: CPU + Chipsets
35
Intel® CPU
Intel®TXT
Intel®SGX
Integrated four pieces into open research platform
Samples & Tools Middleware System Software Hardware Open Research Platform
36
Our personal robot testbeds
SLAM &Navigation navigation 37
FullFull capability interaction
Armless HRI manipulation free
37
Zi-Long: 3D Robot Head for HRI experiments
HW: Core i5 NUC + RealSense R200 + MIC array + Pico projector
38
38
Summary: full-stack HRI research platform Security enhanced Linux/ROS Personal assistant
Kids education
Kids gaming
Tele-medical
Follow me
Patrol
Spatial AR
Home appliance automation
Telepresence
Elderly care
Entertainment
Surveillance
WebRTC
Node-RED SDK
Apps
SDK
Robot SDK Human activity understanding
Emotional intelligence
Scene recognition
People Following/finding/facing
SLAM
Exploration
Environment sensing
ASR
Body Recog.
Object Recog.
V-SLAM
Localization
Navigation
Robot manipulation
TTS
Face Recog..
Text Recog.
Gesture Recog.
Trusted execution Environment Mobile base Wheels Lidar 39
Personalization
Middleware
Virtual Sensor
Capabilities
Algorithms
Middleware
Head
Body UWB
Arms
BLE/WiFi
Camera
RealSense
IMU
Other sensors
Projector /LED
MIC ARRAY
Project or/LED
Trusted Execution Technology
FPGA Acceleration
HW
What will you develop? 40
Legal Notices and Disclaimers Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation.Learn more at intel.com, or from the OEM or retailer. No computer system can be absolutely secure. Tests document performance of components on a particular test, in specific systems.Differences in hardware, software, or configuration will affect actual performance.Consult other sources of information to evaluate performance as you consider your purchase.For more complete information about performance and benchmark results, visit http://www.intel.com/performance. Cost reduction scenarios described are intended as examples of how a given Intel-based product, in the specified circumstances and configurations, may affect future costs and provide cost savings.Circumstances will vary.Intel does not guarantee any costs or cost reduction. This document contains information on products, services and/or processes in development.All information provided here is subject to change without notice.Contact your Intel representative to obtain the latest forecast, schedule, specifications and roadmaps. Statements in this document that refer to Intel’s plans and expectations for the quarter, the year, and the future, are forward-looking statements that involve a number of risks and uncertainties.A detailed discussion of the factors that could affect Intel’s results and plans is included in Intel’s SEC filings, including the annual report on Form 10-K. The products described may contain design defects or errors known as errata which may cause the product to deviate from published specifications.Current characterized errata are available on request. No license (express or implied, by estoppel or otherwise) to any intellectual property rights is granted by this document. Intel does not control or audit third-party benchmark data or the web sites referenced in this document.You should visit the referenced web site and confirm whether referenced data are accurate. Intel, Xeon, Core, RealSense, and the Intel logo are trademarks of Intel Corporation in the United States and other countries.
*Other names and brands may be claimed as the property of others. © 2016 Intel Corporation.
41
Risk Factors The above statements and any others in this document that refer to future plans and expectations are forward-looking statements that involve a number of risks and uncertainties.Words such as "anticipates," "expects," "intends," "goals," "plans," "believes," "seeks," "estimates," "continues," "may," "will," "should," and variations of such words and similar expressions are intended to identify such forward-looking statements.Statements that refer to or are based on projections, uncertain events or assumptions also identify forward-looking statements.Many factors could affect Intel's actual results, and variances from Intel's current expectations regarding such factors could cause actual results to differ materially from those expressed in these forward-looking statements.Intel presently considers the following to be important factors that could cause actual results to differ materially from the company's expectations.Demand for Intel's products is highly variable and could differ from expectations due to factors including changes in business and economic conditions; consumer confidence or income levels; the introduction, availability and market acceptance of Intel's products, products used together with Intel products and competitors' products; competitive and pricing pressures, including actions taken by competitors; supply constraints and other disruptions affecting customers; changes in customer order patterns including order cancellations; and changes in the level of inventory at customers.Intel's gross margin percentage could vary significantly from expectations based on capacity utilization; variations in inventory valuation, including variations related to the timing of qualifying products for sale; changes in revenue levels; segment product mix; the timing and execution of the manufacturing ramp and associated costs; excess or obsolete inventory; changes in unit costs; defects or disruptions in the supply of materials or resources; and product manufacturing quality/yields.Variations in gross margin may also be caused by the timing of Intel product introductions and related expenses, including marketing expenses, and Intel's ability to respond quickly to technological developments and to introduce new products or incorporate new features into existing products, which may result in restructuring and asset impairment charges.Intel's results could be affected by adverse economic, social, political and physical/infrastructure conditions in countries where Intel, its customers or its suppliers operate, including military conflict and other security risks, natural disasters, infrastructure disruptions, health concerns and fluctuations in currency exchange rates.Results may also be affected by the formal or informal imposition by countries of new or revised export and/or import and doing-business regulations, which could be changed without prior notice.Intel operates in highly competitive industries and its operations have high costs that are either fixed or difficult to reduce in the short term.The amount, timing and execution of Intel's stock repurchase program could be affected by changes in Intel's priorities for the use of cash, such as operational spending, capital spending, acquisitions, and as a result of changes to Intel's cash flows or changes in tax laws.Product defects or errata (deviations from published specifications) may adversely impact our expenses, revenues and reputation.Intel's results could be affected by litigation or regulatory matters involving intellectual property, stockholder, consumer, antitrust, disclosure and other issues.An unfavorable ruling could include monetary damages or an injunction prohibiting Intel from manufacturing or selling one or more products, precluding particular business practices, impacting Intel's ability to design its products, or requiring other remedies such as compulsory licensing of intellectual property.Intel's results may be affected by the timing of closing of acquisitions, divestitures and other significant transactions.We completed our acquisition of Altera on December 28, 2015 and risks associated with that acquisition are described in the “Forward Looking Statements” paragraph of Intel’s press release dated June 1, 2015, which risk factors are incorporated by reference herein.A detailed discussion of these and other factors that could affect Intel's results is included in Intel's SEC filings, including the company's most recent reports on Form 10-Q, Form 10-K and earnings release.
Rev. 1/14/16 42