lecture 1 introduction no video
TRANSCRIPT
-
8/14/2019 Lecture 1 Introduction No Video
1/63
Introduction to Autonomous Mobile Robots
Prof. Yan Meng
Department of Electrical and Computer EngineeringStevens Institute of Technology
-
8/14/2019 Lecture 1 Introduction No Video
2/63
Course Logistics
Instructor: Yan Meng
Office: Burchard 411
Phone: 201-216-5496
Email:[email protected]
Office hour: Tuesday 3:00pm-5:00pm
Course website:
http://www.ece.stevens-tech.edu/~ymeng/courses/CPE521/CPE521A.htm
Homework
Homework will be due one week later after it is assigned
Problem solutions will be posted on-line LATE HOMEWORK WILLNOT BE ACCEPTED AFTER THE SOLUTION IS POSTED
Grading Homework 20% Midterm 20% Final 30% Project 30%
mailto:[email protected]://www.ece.stevens-tech.edu/~ymeng/courses/CPE521/CPE521A.htmhttp://www.ece.stevens-tech.edu/~ymeng/courses/CPE521/CPE521A.htmhttp://www.ece.stevens-tech.edu/~ymeng/courses/CPE521/CPE521A.htmhttp://www.ece.stevens-tech.edu/~ymeng/courses/CPE521/CPE521A.htmmailto:[email protected] -
8/14/2019 Lecture 1 Introduction No Video
3/63
Course Syllabus
Required Textbook:
Roland Siegwart and Ilah Nourbakhsh, Introduction to Autonomous MobileRobots, MIT Press, April 2004, ISBN# 0-262-19502-X.
Textbook website: http://autonomousmobilerobots.epfl.ch/
Some reading materials and hands out will be distributed in class.
Recommended readings:
George A. Bekey, Autonomous Robots From Biological Inspiration toImplementation and Control,MIT Press, 2005. ISBN 0-262-02578-7.
Robin Murphy, An Introduction to AI Robotics,MIT Press, November 2000.ISBN 0-262-13383-0.
Stefano Nolfi and Dario Floreano, Evolutionary Robotics: The Biology,Intelligence, and Technology of Self-Organizing Machines, MIT Press,2000, ISBN 0-262-14070-5.
Thomas Braunl, Embedded Robotics: Mobile Robot Design andApplications with Embedded Systems, Springer-Verlag Berlin Heidelberg
New York, ISBN 3-540-03436-6.
-
8/14/2019 Lecture 1 Introduction No Video
4/63
Some Robotics Links
http://www.ifi.unizh.ch/groups/ailab/links/robotic.html#companies
http://www.cooper.edu/~mar/robotics_links.htm
http://www.roboticsonline.com/links/ http://www.ieee-ras.org/
http://www.euronet.nl/users/ragman/link_64.html
http://www.ifi.unizh.ch/groups/ailab/links/robotic.html#companieshttp://www.cooper.edu/~mar/robotics_links.htmhttp://www.roboticsonline.com/links/http://www.ieee-ras.org/http://www.euronet.nl/users/ragman/link_64.htmlhttp://www.euronet.nl/users/ragman/link_64.htmlhttp://www.ieee-ras.org/http://www.roboticsonline.com/links/http://www.cooper.edu/~mar/robotics_links.htmhttp://www.ifi.unizh.ch/groups/ailab/links/robotic.html#companies -
8/14/2019 Lecture 1 Introduction No Video
5/63
Applications of Mobile Robots
Indoor Outdoor
Structured Environments Unstructured Environments
1
-
8/14/2019 Lecture 1 Introduction No Video
6/63
Autonomous Mobile Robots
The three key questions in Mobile Robotics
Where am I ?
Where am I going ?How do I get there ?
To answer these questions the robot has to have a model of the environment (given or autonomously built)
perceive and analyze the environment
find its position within the environmentplan and execute the movement
Basic tasks: deal with Locomotion and Navigation (Perception,
Localization, Planning and motion generation)
1
-
8/14/2019 Lecture 1 Introduction No Video
7/63
-
8/14/2019 Lecture 1 Introduction No Video
8/63
Control Architectures / Strategies
Control Loop
dynamically changing
no compact model available
many sources of uncertainty
1
"Position"Global Map
Perception Motion Control
Cognition
Real WorldEnvironment
Localization
PathEnvironment ModelLocal Map
-
8/14/2019 Lecture 1 Introduction No Video
9/63
Two Approaches
Classical AI(model based navigation)
complete modeling
function based
horizontal
decomposition
New AI(behavior based navigation)
sparse or no modeling
behavior based vertical decomposition
bottom up
Possible Solution Combine Approaches
1
1
-
8/14/2019 Lecture 1 Introduction No Video
10/63
Environment Representation
Continuos Metric -> x,y,
Discrete Metric -> metric grid
Discrete Topological -> topological grid
Environment Modeling
Raw sensor data, e.g. laser range data, grayscale images
o large volume of data, low distinctiveness
o makes use of all acquired information
Low level features, e.g. line other geometric features
o medium volume of data, average distinctiveness
o filters out the useful information, still ambiguities
High level features, e.g. doors, a car, the Eiffel tower
o low volume of data, high distinctiveness
o filters out the useful information, few/no ambiguities, not enough information
Environment Representation and Modeling:
The Key for Autonomous Navigation
1
1
-
8/14/2019 Lecture 1 Introduction No Video
11/63
Odometry
not applicable
Modified
Environments
expensive,
inflexible
Feature-based
Navigation
still a challenge for
artificial systems
Environment Representation and Modeling: How we do it!
Corridorcrossing
Elevator door
Entrance
Eiffel Tower
Landing at nightHow to find a treasure
1
C o u r t e s y K
A r r
a s
1
-
8/14/2019 Lecture 1 Introduction No Video
12/63
Environment Representation: The Map Categories
Recognizable Locations Topological Maps
Metric Topological Maps Fully Metric Maps (continuos ordiscrete)
1
C o u r t e s y K
A r r a s
1
-
8/14/2019 Lecture 1 Introduction No Video
13/63
Incrementally(dead reckoning)
Odometric or initialsensors (gyro)
not applicable
Modifying the environments(artificial landmarks / beacons)
Inductive or optical tracks (AGV)
Reflectors or bar codes
expensive, inflexible
Methods for Navigation: Approaches with Limitations1
C o u r t e s y K
A r r a s
1
-
8/14/2019 Lecture 1 Introduction No Video
14/63
Methods for Localization: The Quantitative Metric Approach
1. A priori Map: Graph, metric
2. Feature Extraction (e.g. line segments)
3. Matching:
Find correspondence
of features
4. Position Estimation:
e.g. Kalman filter, Markov
representation of uncertainties optimal weighting acc. to a priori statistics
1
C o u r t e s y K
A r r a s
1
-
8/14/2019 Lecture 1 Introduction No Video
15/63
Methods for Localization: The Quantitative Topological Approach
1. A priori Map: Graphlocally uniquepoints
edges
2. Method for determiningthe local uniqueness
e.g. striking changes on raw data levelor highly distinctive features
3. Library of driving behaviors
e.g. wall or midline following, blind step,enter door, application specific
behaviorsExample: Video-based navigation withnatural landmarks
Courtesy of [Lanser et al. 1996]
1
1
-
8/14/2019 Lecture 1 Introduction No Video
16/63
Map Building: How to Establish a Map
1. By Hand
2. Automatically: Map Building
The robot learns its environment
Motivation:
- by hand: hard and costly
- dynamically changing environment
- different look due to different perception
3. Basic Requirements of a Map:
a way to incorporate newly sensed
information into the existing world
model information and procedures for
estimating the robots position
information to dopath planning and
othernavigation task(e.g. obstacleavoidance)
Measure of Quality of a map
topological correctness
metrical correctness
But: Most environments are a mixture of
predictable and unpredictable features hybrid approach
model-based vs. behaviour-based
predictability
1
C o u r t e s y K
A r r a s
1
-
8/14/2019 Lecture 1 Introduction No Video
17/63
Map Building: The Problems
1. Map Maintaining: Keeping track ofchanges in the environment
e.g. disappearingcupboard
- e.g. measure of belief of eachenvironment feature
2. Representation andReduction of Uncertainty
position of robot -> position of wall
position of wall -> position of robot
probability densities for feature positions additional exploration strategies
1
C o u r t e s y K
A r r a s
1
-
8/14/2019 Lecture 1 Introduction No Video
18/63
Map Building: Exploration and Graph Construction
1. Exploration
- provides correct topology
- must recognize already visited location
- backtracking for unexplored openings
2. Graph Construction
Where to put the nodes?
Topology-based: at distinctive locations
Metric-based: where features disappear orget visible
1
C
o u r t e s y K
A r r a
s
1
-
8/14/2019 Lecture 1 Introduction No Video
19/63
Automatic Guided Vehicles
Newest generation ofAutomatic Guided
Vehicle of VOLVO usedto transport motorblocks from onassembly station to an
other. It is guided by anelectrical wire installedin the floor but it is alsoable to leave the wire toavoid obstacles. Thereare over 4000 AGV onlyat VOLVOs plants.
1
1
-
8/14/2019 Lecture 1 Introduction No Video
20/63
Helpmate
HELPMATE is a mobile robot used in hospitals
for transportation tasks. It has various on boardsensors for autonomous navigation in thecorridors. The main sensor for localization is acamera looking to the ceiling. It can detect the
lamps on the ceiling as reference (landmark).http://www.ntplx.net/~helpmate/
1
1
-
8/14/2019 Lecture 1 Introduction No Video
21/63
BR700 Cleaning Robot
BR 700 cleaning robotdeveloped and sold byKrcher Inc., Germany
Its navigation system isbased on a verysophisticated sonarsystem and a gyro.http://www.kaercher.de
1
1
-
8/14/2019 Lecture 1 Introduction No Video
22/63
ROV Tiburon Underwater Robot
Picture of robot ROV Tiburon for
underwater archaeology(teleoperated)- used by MBARI fordeep-sea research, this UAV providesautonomous hovering capabilities for
the human operator.
Th Kh R b t1
-
8/14/2019 Lecture 1 Introduction No Video
23/63
The Khepera Robot
KHEPERA is a small mobile robot for research and education. It sizes only about 60mm in diameter. Additional modules with cameras, grippers and much more are
available. More then 700 units have already been sold (end of 1998).http://diwww.epfl.ch/lami/robots/K-family/ K-Team.html
-
8/14/2019 Lecture 1 Introduction No Video
24/63
-
8/14/2019 Lecture 1 Introduction No Video
25/63
SMARbot Overview
CMU cam2
SONAR sensors
Infrared sensors
Bumper switch
Motors withtank treads
Microprocessorboard
FPGA board
Sensor board
Power board
ZigBee wirelessmodule
Forester Robot1
-
8/14/2019 Lecture 1 Introduction No Video
26/63
Forester Robot
Pulstech developedthe first industrial likewalking robot. It is
designed moving woodout of the forest. Theleg coordination isautomated, butnavigation is still doneby the human operatoron the robot.http://www.plustech.fi/
Robots for Tube Inspection1
-
8/14/2019 Lecture 1 Introduction No Video
27/63
Robots for Tube Inspection
HCHER robots for sewage tube
inspection and reparation. Thesesystems are still fully teleoperated.http://www.haechler.ch
EPFL / SEDIREP: Ventilationinspection robot
A t I d N i ti1
-
8/14/2019 Lecture 1 Introduction No Video
28/63
Autonomous Indoor Navigation (Pygmalion EPFL)
very robust on-the-fly
localization
one of the first systemswith probabilistic sensor
fusion
47 steps,78 meter length,
realistic officeenvironment,
conducted 16 times >
1km overall distance
partially difficult
surfaces (laser),
partially few vertical
edges (vision)
Video is here.
SLAM (Si lt l li ti d i ) b
-
8/14/2019 Lecture 1 Introduction No Video
29/63
SLAM (Simultaneous localization and mapping) by
EPFL
M lti b t SLAM ( CMU)
-
8/14/2019 Lecture 1 Introduction No Video
30/63
Multi-robot SLAM ( CMU)
Tour Guide Robot (N b kh h CMU)1
-
8/14/2019 Lecture 1 Introduction No Video
31/63
Tour-Guide Robot (Nourbakhsh, CMU)
Video is here.
Minerva: a second generation museum tour guide robot
-
8/14/2019 Lecture 1 Introduction No Video
32/63
Minerva: a second-generation museum tour-guide robot
Sojourner, First Robot on Mars1
-
8/14/2019 Lecture 1 Introduction No Video
33/63
Sojourner, First Robot on Mars
The mobile robotSojourner was usedduring the Pathfinder
mission to explorethe mars in summer1997. It was nearlyfully teleoperatedfrom earth. However,some on boardsensors allowed forobstacle detection.http://ranier.oact.hq.
nasa.gov/telerobotics_page/telerobotics.shtm
http://www.youtube.com/watch?v=zZWOGcdC_PI
NASA Rover
-
8/14/2019 Lecture 1 Introduction No Video
34/63
NASA Rover
RoboCup 2006
-
8/14/2019 Lecture 1 Introduction No Video
35/63
RoboCup 2006
Midsize Qualification Video Bremen 2006
A short scene from the final in Osaka 05 against Eigen
Modular Robots
-
8/14/2019 Lecture 1 Introduction No Video
36/63
Modular Robots
Modular Reconfigurable Robotics is an approach to building robots for
various complex tasks. Instead of designing a new and different
mechanical robot for each task, you just build many copies of one simple
module. The module can't do much by itself, but when you connect manyof them together you get a system that can do complicated things. In fact,
a modular robot can even reconfigure itself -- change its shape by moving
its modules around -- to meet the demands of different tasks or differentworking environments.
http://www2.parc.com/spl/projects/modrobots/index.html
Self Reconfig rable Robots
-
8/14/2019 Lecture 1 Introduction No Video
37/63
Self-Reconfigurable Robots
Traditional approaches of building separate robots forseparate tasks may not be cost efficient and appropriate forthose complex tasks in environments that are not humanfriendly.
Reconfigurable robot is modular, multifunctional, andreconfigurable for different tasks at different missionstages.
Challenges: how to coordinate all modules to achieve acommon goal dynamically?
Four layers: hardware, locomotion control, transformcontrol, and cognitive control.
Available Reconfigurable Robots MTRAN( National Institute of Advanced Industrial Science
and Technology, Japan)
SuperBot (Polymorphic Robotics Lab, University of SouthernCalifornia)
Molecube (Cornell University)
Others
M-TRAN (Modular Transformer)
-
8/14/2019 Lecture 1 Introduction No Video
38/63
M-TRAN (Modular Transformer)
http://unit.aist.go.jp/is/dsysd/mtran3/
SuperBot (Polymorphic Robotics Laboratory, USC)
-
8/14/2019 Lecture 1 Introduction No Video
39/63
SuperBot (Polymorphic Robotics Laboratory, USC)
http://www.isi.edu/robots/superbot.htm
CrossCube (Stevens Embedded Systems and Robotics Lab)
-
8/14/2019 Lecture 1 Introduction No Video
40/63
CrossCube (Stevens Embedded Systems and Robotics Lab)
Limitations on locomotion designs andhigh-level control algorithms on theavailable reconfigurable robots
Our objective: to tackle those limitationsand develop a highly flexible locomotionmechanism and more intelligent GRN-based cognitive control algorithm to adaptto dynamic environments and tasks.
CrossCube Hardware and locomotion: lattice-based
robot module that is able to rotate, climband parallel move on other modulessurface
Transform and cognitive control: evolvinggene regulation network(GRN) basedalgorithms.
CrossCube
-
8/14/2019 Lecture 1 Introduction No Video
41/63
CrossCube
Self-reconfigure robot modules tovarious shapes/forms based ondifferent task requirements or
environments. Can self-detect module failures and
self-repair malfunctions byreconfiguration
From homogeneous modules toheterogeneous models
Challenges Flexible, robust, adaptive, reliable,
interactive, integration, etc..
Potential applications Urban search and rescue, security,
space exploration, transportationthrough narrow and complex space,etc.
(video demos)
Biological Inspired Robot: Snake Robot (Tokyo Institute
-
8/14/2019 Lecture 1 Introduction No Video
42/63
g p ( y
of Technology, Shigeo Hirose Group)
On the evening of December 26, 1972, for the first time in the world wesucceeded in producing artificial serpentine movement at a speed ofapproximately 40 cm/sec using the principles of a serpentine movementwhich is the same as actual snakes. The entire length of the device is 2 m,
and it has 20 joints. From http://www-robot.mes.titech.ac.jp/robot/snake/acm3/acm3_e.html
Biological Inspired Robot: Snake Robot (Tokyo Institute
-
8/14/2019 Lecture 1 Introduction No Video
43/63
g p ( y
of Technology, Shigeo Hirose Group)
Raise headSerpentine Propulsion
The system consists perpendicularly connected as a straight chain by the unit
that has batteries, a control board, and actuator of 1 DOF, shell structure hadlightweight and high rigidity.
-
8/14/2019 Lecture 1 Introduction No Video
44/63
Biological Inspired Robots: legged robots
-
8/14/2019 Lecture 1 Introduction No Video
45/63
Centralized versus Distributed Control Laws
-
8/14/2019 Lecture 1 Introduction No Video
46/63
Global Centralized Control
Allow for more coherent team cooperation
Often results in increased communication requirements
The knowledge is computationally costly
Oftentimes all the needed global knowledge is not known
Vulnerable with robot failures and in dynamic environment
Local Distributed Control
Computationally Simple
Handle dynamic environments well Oftentimes unclear as to how
to design local control laws
Must rely on physical sensors
Oftentimes unclear as to how to design local control laws
Biological-Inspired Swarm Robots
-
8/14/2019 Lecture 1 Introduction No Video
47/63
Swarm intelligence is an artificial intelligence (AI) technique based on
and modeled after the emergent, decentralized, self-organized,
collective behavior of insect colonies, bird flocks, and animal herds.
SI Natures Design: Insects
-
8/14/2019 Lecture 1 Introduction No Video
48/63
Organizing highways to and
from their foraging sites by
leaving pheromone trails
Form chains from their own
bodies to create a bridge to pull
and hold leafs together with silk
Division of labour between
major and minor ants
SI Natures Design: Birds
-
8/14/2019 Lecture 1 Introduction No Video
49/63
A flight of ducks use V
formation to reduce air drag and
conserve energy Optimize food searches by using
the eyes of other ducks
Ducks in a flight gain
protectionbetter predator
avoidance odds
-
8/14/2019 Lecture 1 Introduction No Video
50/63
Swarm Robots
-
8/14/2019 Lecture 1 Introduction No Video
51/63
Many of the dangerous, dirty, or Null jobs can be performed more effectively by
groups of robots working together, such as swarms.
Applications
Urbane search and rescue,Surveillance systems, Exploration, Constructions Much more .
Advantages
Parallel processing, cover more areas, coordination, robust and flexible
Main challenges
Adapt their behaviors based on interaction with the environment and
other robots
Become more proficient in their tasks over time
Adapt to new situations as they occur
Coordination and cooperation
Swarm-Bots Project ( Marco Dorigo group in Europe)
-
8/14/2019 Lecture 1 Introduction No Video
52/63
The main objective of the Swarm-bots project is to study a novel approach to the
design and implementation of self-organizing and self-assembling artefacts.
This approach was inspired by the recent studies in swarm intelligence in social
insects and other animal societies. An artefact composed of a number of simpler, insect-like, robots, built out of
relatively cheap components, capable of self-assembling and self-organizing to adapt
to its environment
Swarm Robots MIT/iRobot
-
8/14/2019 Lecture 1 Introduction No Video
53/63
http://people.csail.mit.edu/jamesm/swarm.php
Multi-cellular based Multi-Agent Systems (Stevens
Embedded Systems and Robotics Lab)
-
8/14/2019 Lecture 1 Introduction No Video
54/63
Embedded Systems and Robotics Lab)
Self-organization of large collective systems is a challenging task
Autonomous, adaptable, evolvable, robust, self-repairable, emergent
Suboptimal, non-controllable, non-predictable, not (easily) understandable
Trade-off between global (centralized) and local (distributed) control
Biological development, including cell growth, cell differentiation and morphogenesis,
can be seen as a self-organizing process
Robust to genetic and environmental changes
Use of global and local control
Predictable and relatively understandable
Biological development is under the temporal and spatial control of a gene regulatory
network(GRN) Can we borrow some ideas from developmental biology, in particular the
morphogenesis?
60
Simulation Results: Forming shapes
-
8/14/2019 Lecture 1 Introduction No Video
55/63
61The videos can be downloaded from http://www.ece.stevens-tech.edu/~ymeng/lab_home.htm
Preliminary Experimental Results on Multi-Robot
Formation
-
8/14/2019 Lecture 1 Introduction No Video
56/63
Formation
The video demo can be downloaded from http://www.ece.stevens-tech.edu/~ymeng/lab_home.htm
The Honda Walking Robot http://www.honda.co.jp/tech/other/robot.html 1
-
8/14/2019 Lecture 1 Introduction No Video
57/63
http://www.youtube.com/watch?v=kLGk9Q49y7k
Entertainment Robots: Humanoid Robots (SONY)1
-
8/14/2019 Lecture 1 Introduction No Video
58/63
DARPA Grand Challenge
-
8/14/2019 Lecture 1 Introduction No Video
59/63
The DARPA Grand Challenge has been the most significant event for
the robotics community in more than a decade.
A mobile ground robot had to traverse 132 miles of unrehearsed desert
terrain in less than 10 hours.
In 2004, the best robot only made 7.3 miles.
In 2005, Stanford won the challenge and the $2M prize in less than 7hours travel time, and ahead of four other finishers.
Stanford STANLEY (http://robots.stanford.edu/)
-
8/14/2019 Lecture 1 Introduction No Video
60/63
DARPR Grand Challenge 2007
-
8/14/2019 Lecture 1 Introduction No Video
61/63
The Urban Challenge. Teams will compete to build an autonomousvehicle able to complete a 60-mile urban course safely in less than 6hours.
The DARPA Urban Challenge will take place in Victorville, Californiaon November 3, 2007.
"It was an important step to have autonomous ground vehicles thatcan navigate and drive across open and difficult terrain from city tocity. But the next big leap will be an autonomous vehicle that cannavigate and operate in traffic, a far more complex challenge for a
'robotic' driver. So this November we are very excited to be movingfrom the desert to the city with our Urban Challenge."
Dr. Tony Tether, Director, DARPA
Unmanned Maritime System (Senior Design project)
-
8/14/2019 Lecture 1 Introduction No Video
62/63
Point of Contact: M. DeLorme (Center for Maritime Systems)
No of Students: 2
Fields of Interest: Robotics, autonomous systems
Project Sponsor: Office of Naval Research
DESCRIPTION:
The project involves the design, development and demonstration deployment of an unmannedmaritime system (UMS) or systems to perform a task to be specified by the project sponsor.
Students will be responsible for developing the system and deployment specifications based
on independent research and planning. This team will be part of a larger multidisciplinary
team working with students in Mechanical Engineering and Naval Engineering to accomplishthe project goals. Interested students MUST meet with Michael DeLorme
([email protected]) to further discuss the responsibilities and expectations of this project
and to submit a one page resume highlighting their qualifications as related to the proposed
project.
Homework #1
-
8/14/2019 Lecture 1 Introduction No Video
63/63
In order to prepare your project, you may want to search for some
robot simulators from the websites. Please try to find at least two
robot simulators you like and try to use them to see if it is possible for
you to write control programs, such as localization, navigation, multi-robot coordination, on those simulators.
You can find your project partners and build up a group (at most 3
persons for undergraduates, and 2 persons for graduates), or you like todo it individually (more credits).
For the course project, you have two options
Theoretical exploration: real research papers and propose some newapproaches
Building robotic systems, which includes building real robotic systems or
running on a simulation