16-899C Statistical
Techniques In Robotics
Sebastian Thrun and Geoffrey Gordon
Carnegie Mellon University
www.cs.cmu.edu/~thrun
www.cs.cmu.edu/~ggordon
© sebastian thrun, CMU, 2000
1
notes
 Pointer to Larry’s material
© sebastian thrun, CMU, 2000
2
Administrative Information








Sebastian Thrun
Geoffrey Gordon
Web:
Email list:
Time:
Location:
TA:
Appointments:
thrun@cs.cmu.edu
ggordon@cs.cmu.edu
http://www.cs.cmu.edu/~16899
16-899@cs.cmu.edu
Mon/Wed, 10:30-11:50am
NSH 3302
n/a
send Email!
© sebastian thrun, CMU, 2000
3
© sebastian thrun, CMU, 2000
4
Goals
 Enable you to program robots and embedded systems in a
robust fashion
 Enable you to understand the intrinsic assumptions in your
robot software
 Enable you to pursue original research in probabilistic
robotics
 Sway you into joining a young and fascinating research
field: probabilistic robotics
© sebastian thrun, CMU, 2000
5
What this course is not
 Intro to robotics
 Little work
 Low on math
© sebastian thrun, CMU, 2000
6
Course Schedule
Localization
Sept 4-16
Mapping
Sept 30-Oct 16
Decision Making
Oct 21-30
Multi-Agent
Nov 4-11
Advanced Perception
Nov 13-25
© sebastian thrun, CMU, 2000
7
What You Should Do
 Think
 Think differently
 Be critical
 Come up with Original Research
© sebastian thrun, CMU, 2000
8
What Is A Good Project
 Mine Mapping
 Multi-Agent Control
© sebastian thrun, CMU, 2000
9
Requirements
 In teams of three:
• Warm-up project (mobile robot localization)
• Written assignment(s)
• Research Project
 Class Presence: all but two sessions (send me email)
 Quizzes (all but at most two)
 No exams
© sebastian thrun, CMU, 2000
10
Your next tasks
 Check out Web site
• Read assigned paper
• Download map+sensor data and program robot
localization algorithm
 Send Sebastian mail with your name and names
of team mates (for warm-up project)
 Come to class on Sept 9th (10:30am-11:50am)
© sebastian thrun, CMU, 2000
11
© sebastian thrun, CMU, 2000
12
© sebastian thrun, CMU, 2000
13
© sebastian thrun, CMU, 2000
14
© sebastian thrun, CMU, 2000
15
Five Sources of Uncertainty
Approximate
Computation
Environment
Dynamics
Random
Action Effects
Inaccurate
Models
Sensor
Limitations
© sebastian thrun, CMU, 2000
16
Trends in Robotics
Classical Robotics (mid-70’s)
• exact models
• no sensing necessary
Reactive Paradigm (mid-80’s)
• no models
• relies heavily on good sensing
Hybrids (since 90’s)
• model-based at higher levels
• reactive at lower levels
Probabilistic Robotics (since mid-90’s)
• seamless integration of models and sensing
• inaccurate models, inaccurate sensors
© sebastian thrun, CMU, 2000
17
© sebastian thrun, CMU, 2000
18
Rhino
© sebastian thrun, CMU, 2000
19
Minerva
© sebastian thrun, CMU, 2000
20
The CMU/Pitt Nursebot Initiative
© sebastian thrun, CMU, 2000
21
People Detection
Mike Montemerlo
© sebastian thrun, CMU, 2000
22
Learning Models of People
Maren Bennewitz
© sebastian thrun, CMU, 2000
23
3D Mapping Result
With: Christian Martin
© sebastian thrun, CMU, 2000
24
Multi-Robot Exploration
© sebastian thrun, CMU, 2000
25
Mine Mapping (brand new)
© sebastian thrun, CMU, 2000
26
What are interesting problems?




Mapping, automatic, manual, guided?
Probabilistic localization, landmarks?, odometer!,
Route planning, collision avoidance
Mine Mapping?
© sebastian thrun, CMU, 2000
27
How can we solve them?
© sebastian thrun, CMU, 2000
28
© sebastian thrun, CMU, 2000
29
Where Am I/?
© sebastian thrun, CMU, 2000
30
Nature of Sensor Data: Uncertainty
Odometry Data
© sebastian thrun, CMU, 2000
Range Data
31
© sebastian thrun, CMU, 2000
32
Warm-Up Assignment: Localization,
Due Sept 23
© sebastian thrun, CMU, 2000
33
Warm-Up Assignment: Localization
© sebastian thrun, CMU, 2000
34
Warm-Up Assignment: Localization
© sebastian thrun, CMU, 2000
35