Lecture 1 - Introduction Flashcards

1
Q

What is a robot (4)?

A

A robot is an autonomous system (1) which exists in the physical world (2, embodiment), can sense its environment (3, situated), and can act on it to achieve goals (4).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Autonomous

A

Acting, based on its own decisions, and not controlled by a human

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Teleoperated

A

Opposite to autonomous: externally controlled by humans

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Autonomous System (2)

A

Capable to maintain itself (1) in the outside world and to adapt and react differently (2) to the changing environment (extreme case learning mechanism).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Embodied

A

a body, being able to act in the world

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Situated

A

being able to perceive the world

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Sensors (2) and Actuators (2) of Tortoises (early robots)

A

Sensors: Light and touch
Actuators: Forwards/Backwards and Steering.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Braitenberg Vehicle 2a Fear

A

Straight, excitatory connection

Avoids light source and moves slower when further away

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Braitenberg Vehicle 2b Aggression

A

Crossed, excitatory connection

Approaching the source, moving faster when approaching.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Frame-of-reference problem

A

The observer assigns a motivation to the agent, which is not the case.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Example Frame-of-reference problem:

3 aggression vehicles, two with lights

A

Emergent behavior: When two agents are stuck against the wall, the third helps.
Actual behavior: The third agent is strongly attracted by the two lights of the stuck agents.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Braitenberg Vehicle 3a ‘Love’

A

Straight, inhibitory connections

Approaches light source and moves slower when approaching it.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Braitenberg Vehicle 3b ‘Explorer’

A

Crossed, inhibitory connections

Avoids light source and moves slower close to it.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Braitenberg Vehicle 3c

A

Multi-sensorial with sensors for different qualities with varied excitatory, inhibitory, crossed or non-crossed connections.
Attracted by light, but turns away when close as it avoids obstacles.
-> different behavior without action selection mechanism.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Braitenberg Vehicle 4 ‘Values and Special Tastes’-

A

Has non-linear activation function

-> increase in signal does not produce proportional in- or decrease in motor speed (it is non-linear)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Braitenberg Vehicle 4a

A

Continuous but non-linear activation function.
Behavior: 1. Approaching light source (moving faster when closer)
2. At certain point, moving slower and turning from light source.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Braitenberg Vehicle 4b

A

Discontinuous activation function, e.g. has a threshold

-> Agent sits still and suddenly starts to operate

18
Q

What do Braitenberg Vehicles 1-4 have in common (4)?

A
  1. Purely reactive systems
  2. No learning involved, no history.
  3. Behavior is complex and hard to predict (especially with more sensors)
  4. Limited level of autonomy.
19
Q

Braitenberg Vehicle 5 ‘Logic’

A

Threshold devices with logical Nodes which includes a hidden layer

20
Q

Recurrent logical connections equal to what?

A

Memory

21
Q

Segmentation of behavior (example vehicle 3c)

A

Observers segment behavior: first turning to light, then action of turning away.
Actual behavior: result of the internal process in interaction with environment. There is no action selection!

22
Q
Downhill invention (1) + Uphill analysis (2)
-> Which approach do they support (3)?
A

(1) Easy to build vehicles and increase complexity
(2) Difficult to understand why the robot exhibits certain behavior
(3) Supports synthetic approach (“Understanding by building”).

23
Q

What does complexity of behavior depend on (2)?

A

(1) Complexity of robot

(2) Complexity of world

24
Q

Simulated robots

(1) advantages
(2) focus on

A
(1)
- No hardware problems
- Simple observations
(2)
Collective behavior
25
Q

Reynold’s Boids

  • Rules (3)
  • > leads to
A

(1) Close: Don’t collide
(2) Middle: Same direction
(3) Far away: group!
- > leads to realistic group behavior

26
Q

Boids, tortoises and Braitenberg vehicles mimic biological creatures and therefore are…
+ What kind of control do they have?

A

Biomimetic robots

+They have reactive control

27
Q

Built in reflexes leads to behavior that is not built in. This is called:

A

Emergent behavior.

28
Q

Photophilic

A

“loving light” - some of the Braitenberg vehicles

29
Q

Photophobic

A

“afraid of light” - Braitenberg

30
Q

Which fields did robots grow out of (3)?

A

Control theory (mathematics of controlling machines), cybernetics (integration of sensing, action, and environment), and AI (mechanisms for planning and reasoning).

31
Q

(4) Main components of a robot

A
  1. A physical body (to exist)
  2. Sensors (to perceive environment)
  3. Effectors and actuators (to take actions)
  4. A controller (to be autonomous)
32
Q

State space

+ example for discrete and continuous

A

All possibles states a systems can be in.
Ex.:
1. The discrete Light switch has two discrete states (on or off)
2. The dimmable light switch is continuous.

33
Q

Representation / Internal model

A

The internal state used to remember information about the world (e.g. path through a maze).

34
Q

Sensor/perceptual space

A

Space of all possible sensory readings

35
Q

Effectors

A

Enables robot to take action. Similar to legs, wings or other body parts. They need underlying mechanisms (actuators) to work.

36
Q

Actuators

A

Do the actual work and enable action. Similar to muscles.

37
Q

Effectors and actuators are used for (2) main activities…

A
  1. Locomotion: moving around

2. Manipulation: handling objects

38
Q

2 major subfields of robotics

1. Mobile robotics

A

Concerned with robots that move around mostly on the ground but also in the air and under water.

39
Q

2 major subfields of robotics

2. Manipulator robotics

A

Robotic arms of various types

40
Q

Controllers

A

provide the hardware and/or software that makes the robot autonomous by using the sensor inputs and any other information (memory) to decide what to do.