networked robotic systems - auburn university
TRANSCRIPT
CRR Lab - A Brief History
• Outgrowth of work started at Eglin AFB in 1992
• Infrared / Millimeter-Wave Radar Sensor Fusion
• Follow-on funding DARPA e-NOSE
• Best sensor platform? Robots• Many robots are better than one robot..
SENSOR FUSION LABORATORY
Problem Complexity: Human vs. Machine
HUMAN
MAC
HIN
E
EASY HARD
EASY
HARDMaximum Potential Benefit
• Object recognition• Linguistics• Extraction of Relevant Features
from Sensor Arrays
• Arithmetic• Logic
• Thresholding• Tallying
• Judging
IR / MMW DATA FUSIONSupport: AFOSR 1992-93
Project Goal: Improved identification of military vehicles from aerial scenes.
LANCE Missile Launcher
T-62 Tank
M-113 Armored Personnel Carrier (APC)
IR / MMW Fusion, cont’dAPPROACH:
IR SCENE PIXELS
MMW RADAR DATA
NEURAL NETWORK
APCTANKLAUNCHER
PERFORMANCE ASSESSMENT: A T LA + - -T - + -L - - +
• Multiple permutations
• Confusion matrix
• Average result
OVERALL RESULT: 14 % improvement with sensor fusion
Chemical Sensor ArraysSupport: DARPA 1997-99
PROJECT GOAL: Improved identification and detection of chemical plumes in non-laboratory conditions.
VEHICLE
SENSORS
PLUME COMMAND
STATIONRF LINK
ROAD
WIND
Canine Training at IBDSAuburn is world-renowned for training of detection dogs at the Institute for Biological Detection Systems.
Chemical Sensor Arrays, cont’d
Odor Sensor Array
0 100 200 300 400 5000
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Timestep
Sen
sor V
olta
ge
Sensor Outputs
Sensor Array Dynamic Response
Chemical Sensor Arrays, cont’d
0 100 200 300 400 5000
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Timestep
Sen
sor V
olta
ge
10 20 30 40 50
2468
101214
Sen
sor N
umbe
rTimestep
Sensors 1-15
Raw Output Thresholded Binary Output
Above ThresholdBelow ThresholdPreprocessing
Chemical Sensor Arrays, cont’d
ace
Sample 1 Sample 21
20
Sample 31
20
amm
dal
g87
g89
g93
oil
pth
Sensor #
xyl
5 10 15Sensor #5 10 15
Sensor #5 10 15
Chemical Sensor Arrays, cont’d
input categories
netw
ork
resp
onse 1 timestep
aceammdalg87g89g93oilpthxyl
5 timesteps 10 timestepsne
twor
k re
spon
se 20 timestepsaceammdalg87g89g93oilpthxyl
50 timesteps Ideal Response
Time Evolution of Confusion Matrix: Forward SequenceTrained for 20 timesteps
00.10.20.30.40.50.60.70.80.91
Chemical Sensor Arrays, cont’d
00.10.20.30.40.50.60.70.80.91
Time Evolution of Confusion Matrix: Random SequenceTrained for 20 timesteps
1 timestep 5 timesteps 10 timesteps
20 timesteps 50 timesteps Ideal Response
netw
ork
resp
onse
aceammdalg87g89g93oilpthxyl
netw
ork
resp
onse
aceammdalg87g89g93oilpthxyl
input categories
BIOMIMETICSSupport: Under discussion with AF Advanced Guidance Division, Munitions Directorate at Eglin AFB
PROJECT GOAL: Learn sensor fusion from animals. Apply this to flying a drone to target using onboard video.
Flies land accurately
Bees find flowers
Bats catch evading insects in flight
CRR Lab – History, Cont’d
• Feb. 2006: Invited Joe Albree – Math Prof. at AUM - to speak to HKN about history of the engineering profession in USA.
• I didn’t know he co-authored a book about the history of West Point with…
• Gen. Chris Arney, ARO program in Multi-Agent Systems, who was organizing…
• LIMES 2006 at West Point. Language for Intelligent Machines.
Cooperative Autonomous Robots for ReconnaissanceWhite Paper for Chris Arney, AROPrepared 8/29/2005 by Thad Roppel, ECE Dept., Auburn UniversityContact: [email protected], (334) 844-1814
Eric Hildebrand ELEC 5530 HW 4 November 10, 2010
Dominion
The year was 2143, and humanity was at the will of a single man. Known only as “Roppeth”, an evil mastermind had created an army that defeated everything humanity threw at it. No one knew where this army came from, but it could only be assumed that Roppeth created the first generation, and each new generation was spawned by the previous. What made the army so overpowering was the fact that they were autonomous robots controlled by the will of their leader but could act and behave independently from his control. These robots were bipeds, slightly larger than an average human, but completely overpowering to any human counterpart. …
ASIMO
• Highly functional biped
• The future…?
• Video
• But for now, cooperation is more like this….
Oct. 2008 – Robotics and Autonomous Systems - Special Issue on Network Robot Systems
• A probabilistic framework for entire WSN localization using a mobile robot
• Action evaluation for mobile robot global localization in cooperative environments
• Autonomous functional configuration of a network robot system
• Framework and service allocation for network robot platform and execution of interdependent services
• Robots in the kitchen: Exploiting ubiquitous sensing and actuation
• Human behavior recognition using unconscious cameras and a visible robot in a network robot system
• End-to-end congestion control protocols for remote programming of robots, using heterogeneous networks: A comparative analysis
NRS Definition• The IEEE Society of Robotics and Automation Technical Committee on Networked
Robots provides the following definition of Networked Robots
• Physical embodiment: Any NRS has to have at least a physical robot which incorporates hardware and software capabilities
• Autonomous capabilities: A physical robot must have autonomous capabilities to be considered as a basic element of a NRS.
• Network-based cooperation: The robots, environment sensors and humans must communicate and cooperate through a network.
• Environment sensors and actuators: Besides the sensors of the robots, the environment must include other sensors, such as vision cameras and laser range finders, and other actuators, such as speakers and switches
• Human-robot interaction: In order to consider a system as NRS, the system must have a human-robot related activity.
NRS Definition Expanded
Two subclasses of Networked Robots:
(1) Tele-operated*human supervisors send commands and receive feedback via the network.
-Medicine, education, search & rescue,…
(2) Autonomous, *robots and sensors exchange data via the network.
*sensor network extends the effective sensing range of the robots
*allows them to communicate with each other over long distances to coordinate their activity.
*The robots in turn can deploy, repair, and maintain the sensor network to increase its longevity, and utility.
*Broad challenge: develop a science base that couples communication to control to enable such new capabilities
Network Robot Types
Three types of network robots: • Visible - can be seen
– humanoid, pet, stuffed animal, etc.• Virtual - acts in a cyber space and makes use of
information available on Internet. – avatar agent on a mobile phone or info kiosk
• Unconscious - users are not aware of the presence of the robot– camera or a sensor embedded in
infrastructure
Japan NRSJapan NRS consists of four major Japanese companies:• NTT - telecommunications; •Toshiba - home appliances; •Mitsubishi Heavy Industries - industrial robots•ATR- telecommunication and social robotics R&D
NRS in USA
• NetBot Lab at TAMU (Prof. Dezhen Song)
• Ghostrider video clip
• DARPA, JPL
• Georgia Tech
WSN Localization, cont’d
(a)Scheme of the approach. The signal strength is used to estimate the position of the nodes of the network. The mobile robot computes centrally an initial estimation employing a separate Particle Filter for each node. In the second step, a decentralized Information Filter integrates information received from neighbor nodes and the robot, at each node.
(b): An example, a ground robot (Romeo) driving through the network.
ROS (Willow Garage)
• Willow Garage