Download presentation
Presentation is loading. Please wait.
1
Mobile ROBOTS – control systems
Doc. Dr. Ing. Tomáš Brandejský FD ČVUT, k614, k620, LAI k302
2
The first mobile robots
Program versus asyncronnous events Commonly, it is impossible to propose all events which might happen, their possible sequences Thus, it is difficult to create good program suitable for mobile robot which shall move in common, non-model, unknown environment
3
M První mobilní roboti
4
M The first robots
5
Common sense knowledge
Designer of control system has a large knowledge about real environment and relations, which are valid in it. They are so typical, that he does not reason missing of this knowledge in robot There is a big amount of these common sense knowledge
6
The first mobile robots
SMPA Continues in good old fashion artificial intelligence It was very slow on computers of the last 70th Also expensive to memory for model representation
7
Model consistence problem
SMPA Who is behind the doors? What is behind of suitcase and how this suitace looks from bottom? What is it moves (car, branch)? How is model validity limited in time? Exist complete model? Exist consistent model?
8
Reprezentation of environment model
SMPA Relations – geometrical, temporal (kauzal), features and behaviors There are also different possibilities of representation Rules Objects Frames ....
9
The best model of environment is environment
SRA Rodney Brooks Intelligence without reason Intelligence without representation Animals modelling The most primitive beings are not able to produce model of environment, but thir behavior is “intelligent”
10
M R. Brooks
11
Unconditional reflexes
M Unconditional reflexes SRA SRA requires sensors Relation between intelligence and data flow from sensors Paralel reactions What if these reactions are contradictory? (latter)
12
Unconditional reflexes
M Unconditional reflexes
13
M Conditional reflexes
14
Conditional reflexes SRA I. P. Pavlov and his dog Machine lerning
Representation determines, what might be described, learned!
15
Parallel stimuli SRA Fusion of parallel reactions Priorities Fuzzy ANN
Composition respecting stimuli (asociativní)
16
Subsuption architecture
M Subsuption architecture Evolution of control system Providing units with sensors and actuators Self-organization Communication system as “table” ? Efficiency ? order of stimuli
17
Subsuption architecture
M Subsuption architecture
18
Subsuption architecture
M Subsuption architecture
19
Simulation Simulation is not able to replace reality
Everytime it is simplification Also model might ommit some “common sence” feature Simulation system OtoSim
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.