Download 2015_0042_Quantum Robot = CSP = Quantum Emotional

Quantum
Robots
QUANTUM
ROBOTICS IN
ROBOT THEATRE
Quantum Signals and Automata
Binary Logic
0, 1
Fuzzy Logic
[0,1]
Quantum
Logic
Hilbert
Space,
Bloch
Sphere
Quantum Signals and Automata
Logic circuit
Quantum
array
Finite State
Machine
Quantum state
Machine
Algorithm
Quantum
Algorithm
A “Quantum Robot” Concept
Quantum
Computing
Since
1999
Quantum
Robotics
Intelligent
Robotics
Since
2004
Since
1969
Quantum
Robotics
Constraint Satisfaction
Model for Grover
Algorithm
Collaboration:
Martin Lukac, Michitaka Kameyama,
Tohoku University,
Vamsi Parasa, Erik Paul
Quantum Robot Vision
Quantum Fuzzy Logic
Arushi Raghuvanshi
Michael Miller, Univ. Victoria, BC
Quantum Braitenberg
Vehicles
Quantum Emotions
Quantum Initialization
and Neural Networks
Siddhar Manoj
Arushi
Raghuvanshi
Martin Lukac
David Rosenbaum
The first quantum
robot in the world?
Our concept of
quantum robot
based on reducing all
problems to constraint
satisfaction solved on a
quantum computer
Orion Quantum
Adiabatic
Computer in
Vancouver BC,
Canada
Orion
interface
Personal
computer
PC
Bluetooth
connection
CUDA/GPU supercomputer
The whole proposed PSU Quantum
Robot system
QUBOT-1 – the
world’s first
quantum robot
Constraints Satisfaction
Problems
SEND
+MORE
MONEY
Graph coloring
Cryptographic
Problems
Grover Algorithm
Reminder in new light
Graph Coloring
•
•
Building oracle for graph coloring is a better explanation of Grover than
database search.
This is not an optimal way to do graph coloring but explains well the principle
of building oracles.
The Graph Coloring Problem
1
3
2
5
4
Color every node
with a color. Every
two nodes that share
an edge should have
different colors.
Number of colors
should be minimum
1
3
2
5
4
6
6
7
7
This graph is 3-colorable
Simpler Graph Coloring Problem
1
3
2
We need to give all
possible colors
here
4
Two wires for color of node 1
Two wires for color of node 2
Two wires for color of node 3
Two wires for color of node 4

Gives “1” when nodes 1 and
2 have different colors

12

13

23
24

34
F(x)
Value 1 for good coloring
Oracle for Quantum Map of Europe Coloring
Germany
France
Switzerland
Spain
Spain
France
Germany
Switzerland
MultipleValued
Quantum
Circuits




Quaternary
qudits
Good
coloring
Constraint Satisfaction
Problem is to satisfy the
constraints and minimize the
energy
Germany
France
Switzerland
Spain
Spain
France
Constraints to
be satisfied
Germany
Switzerland




Good
coloring
Energy to be
minimized
Count number of ones
Adiabatic Quantum
Computer
Simpler Graph Coloring Problem
Give
Hadamard for
each wire to
get
superposition
of all state,
which means
the set of all
colorings
We need to give all
possible colors
here
|0>
|0>
|0>
H
H
H
H
H


12
Discuss naïve nonquantum circuit with
a full counter of
minterms

13

23
24

34
f(x)
Value 1 for good coloring
Now we will generate whole Kmap at once
using quantum properties - Hadamard
What Grover algorithm does?
•
Grover algorithm looks to a very big Kmap and tells where is the -1 in it.
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
-1 1
1 1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Here
is -1
“Classical” Quantum Computer Circuit
Model for Graph Coloring
Sorting/
Absorber
Color #
Counter
inputs
C
C
AND
Desired
number of
colors
Compar
ator
AND
Output
Simplified schematic of our Graph Coloring Oracle.
We designed 35 oracles
Hadamard Transform
Single qubit

=
1/2
=
1
1
1
1
1
-1
1
-1
1 1
-1
-1
1 -1
-1
1
Here I calculated Kronecker
product of two Hadamards
H
H
H
Parallel connection of
two Hadamard gates is
calculated by Kronecker
Product (tensor product)
• As we remember, these are transformations of Hadamard gate:
Motivating calculations for 3 variables
|0>
H
|0> + |1>
|1>
H
|0> - |1>
In general:
|x>
H
|0> + (-1) x |1>
For 3 bits, vector of 3 Hadamards works as follows:
|abc> 
(|0>+(-1)a|1>) (|0>+(-1)b|1>) (|0>+(-1)c|1>) =
From
multiplication
|000> +(-1)c |001> +(-1)b |001>+(-1)b+c |001>000> +(-1)a |001> +
(-1)a+c |001> + (-1)a+b |001> (-1)a+b+c |001>
If a = b = c =0 then all
phases positive
|0>
oracle
|000> +|001> +
|010>+|011>
+|100> + |101>
+|110> + |111>
|0>
We can say that Hadamard
gates before the oracle create
the Kmap of the function,
setting the function in each of
its possible minterms (cells) in
parallel
This is like a Kmap with every
true minterm (1) encoded by -1
And every false minterm (0)
encoded by 1
f(x)
Block Diagram for graph coloring and
similar problems
Vector
Of
Basic
States
|0>
All good colorings are encoded
by negative phase
Vector
Of
Hadamards
Vector
Of
Hadamards
Oracle with Comparators,
Global AND gate
Think about
this as a very
big Kmap with
-1 for every
good coloring
Output of
oracle
Work
bits
A practical Example
• This presentation shows clearly how to perform
a so called 1 in 4 search
• We start out with the basics
1 in 4 search
Pick your needle and I will find you a haystack
The point of this slide is to show
examples of 4 different oracles.
Grovers search can tell between
these oracles in a single iteration,
classically we would need 3
iterations.
Properties of the oracle
Let f : {0,1}2  {0,1} have the property that there is exactly
one x  {0,1}2 for which f (x) = 1
Goal: find x  {0,1}2 for which f (x) = 1
Classically: 3 queries are necessary
Quantumly: ?
Only after 3 tests can we
determine with certainty
that the oracles is 1 for
only a single input value x
A 1-4 search can chose between 4 oracles in
one iteration
Black box for 1-4 search:
x1
x1
f
x2
x2
y
y  f(x1,x2)
Start by creating phases in superposition of all inputs to f:
0
0
1
H
H
H
f
Input state to query:
(00 + 01 + 10 + 11)(0 – 1)
Output state:
((–1) f(00)00 + (–1) f(01)01 + (–1) f(10)10 + (–1) f(11)11)(0 – 1)
Here we clearly see the Kmap encoded in phase – the main property of many quantum
algorithms
This slide illustrates how the state of the system is changed as it propagates
through the quantum network implementation of Grovers Search algorithm.
Time
0
0
H
H
1
H
state =
0
1
0
0
0
0
0
0
ab c
00
01
11
10
01
1
f
state =
0.353
-0.353
0.353
-0.353
0.353
-0.353
0.353
-0.353
ab c
01
0.3 –0,3
00
01 0.3 –0,3
11 0.3 –0,3
10 0.3 –0,3
X
X
H
H
state =
0.353
-0.353
0.353
-0.353
0.353
-0.353
-0.353
0.353
ab c
state =
0.353
-0.353
0.353
-0.353
0.353
-0.353
-0.353
0.353
01
00 0.3 –0,3
01 0.3 –0,3
11 - 0.3 0,3
10 0.3 –0,3
H
state =
-0.353
0.353
0.353
-0.353
0.353
-0.353
0.353
-0.353
ab c
01
H
state =
0
0
-0.5
0.5
0.5
-0.5
0
0
X
X
H
H
M
M
H
M
state =
0
0
-0.5
0.5
0
0
0.5
-0.5
ab c
01
00 0.3 –0,3 00 - 0.3 0,3
01 0.3 –0,3 01 0.3 –0,3
11 - 0.3 0,3 11 0.3 - 0,3
10 0.3 –0,3 10 0.3 – 0,3
ab c
01
ab c
01
00 0 0 00 0
01 - 0.5 0,5 01 - 0.5
11 0 0 11 0.5
10 0.5 – 0,5 10 0
0
0,5
- 0.5
0
Time
0
0
H
H
1
H
f
X
X
H
H
state =
0.353
-0.353
0.353
-0.353
0.353
-0.353
-0.353
0.353
ab c
01
Ibverters flip between
00 and 11
ab c
01
00 0.3 –0,3 00 - 0.3 0,3
01 0.3 –0,3 01 0.3 –0,3
11 - 0.3 0,3 11 0.3 - 0,3
10 0.3 –0,3 10 0.3 – 0,3
Hadamard addis in
00 and 11
ab c
01
H
state =
-0.353
0.353
0.353
-0.353
0.353
-0.353
0.353
-0.353
H
state =
0
0
-0.5
0.5
0.5
-0.5
0
0
01
00 0 0 00 0
01 - 0.5 0,5 01 - 0.5
11 0 0 11 0.5
10 0.5 – 0,5 10 0
0
0,5
- 0.5
0
state =
-0.353
0.353
0.353
-0.353
0.353
-0.353
-0.353
0.353
state =
0
0
-0.5
0.5
0
0
0.5
-0.5
ab c
01
00 -0.3
01 0.3
11 -0.3
10 0.3
0.3
-0.3
0.3
-0.3
H
H
M
M
H
M
state =
0
0
0
0
0
0
0
-1
state =
-0.353
0.353
0.353
-0.353
0.353
-0.353
-0.353
0.353
Ibverters flip between
00 and 11
Inverter flips second
bit when first is 1
ab c
X
X
ab c
state =
0
0
0
0
0
0
0
1
Hadamard of
affine function
01
00 - 0.3 0.3
01 0.3 - 0.3
11 - 0.3 0.3
10 0.3 - 0.3
ab c
00
01
11
10
01
-1
Grover Loop
Time
0
0
H
H
1
H
f
H
H
X
X
H
H
X
X
H
H
M
M
H
M
Inversion about the mean
ψ00 = – 00 + 01 + 10 + 11
ψ01 = + 00 – 01 + 10 + 11
ψ10 = + 00 + 01 – 10 + 11
ψ11 = + 00 + 01 + 10 – 11
The state corresponding to the
input to the oracle that has a
output result of 1 is ‘tagged’ with
a negative 1.
We need to repeat the Grover
Loop N times
After Hadamard
the solution is
“known” in Hilbert
space by having
value -1. But it is
hidden from us
This was a special case where we could
transform the state vector without
repeating the oracle.
In general we have to repeat the oracle –
general Grover Loop
Measurements
Hadamards
Constants
Grover
Loop
Quantum Braitenberg
Vehicle
LEARNING
Outputs actuators
Measurements
Grover
Loop
Worst case quadratic speedup on
every problem that you can build an
oracle!
Outputs actuators
Controlled
Hadamards
Constants
Inputssensors
Control
Measurements
Inputssensors
Oracle or
Quantum
Circuit
Grover Search
New
Concept of
Real-time
Quantum
Search
Future work
Our concept of
quantum robot
based on reducing all
problems to constraint
satisfaction solved on a
quantum computer
Robot
Obstacle
Avoidance
Problem
Robot
Communication
Problem
Constraint Satisfaction Problem
Robot Reasoning
Problem
Robot Vision
Problem
New Approach
to Quantum
Robotics
Adiabatic Quantum Computer
Classical quantum
computing
Orion Quantum
Adiabatic
Computer in
Vancouver BC,
Canada
Complete multiprocessor
classical robot
Real
quantum
computer
Orion
interface
Personal
computer
PC
CUDA/GPU supercomputer
The whole proposed PSU Quantum
Robot system
Powerful quantum computer
simulator
SAT as a constraint satisfaction
problem
(a + b’ + c) * (b + d’) … = 1
=1
=1
Highly parallel
system of
updating nodes
c
a
=0
=1
b
d
(a + b’ + c)
(b + d’)
Yes , do nothing
No , update
to nodes
nodes
SAT as a constraint satisfaction problem
(a + b’ + c) * (b + d’) … = 1
Constraints:
(a + b’ + c) = 1
(b + d’) = 1
…..
Energy optimization:
(a + b’ + c) = f1
(b + d’) = f2
….
Min ( f1’ + f2’ + …..)
Orion
programming
is just writing
equations for
constraints
and
equations for
energy
Constraint Satisfaction for Robotics
• Insufficient speed of robot image processing and pattern
recognition.
– This can be solved by special processors, DSP processors, FPGA
architectures and parallel computing.
• Prolog allows to write CSP programs very quickly.
• An interesting approach is to formulate many problems using the
same general model.
• This model may be predicate calculus, Satisfiability, Artificial
Neural Nets or Constraints Satisfaction Model.
– Constraints to be satisfied (complex formulas in general)
– Energies to be minimized (complex formulas)
Constraint satisfaction model in robotics
Used in main areas of robotics:
– vision,
– knowledge acquisition,
– knowledge usage.
• In particular the following:
– planning, scheduling, allocation, motion planning, gesture planning,
assembly planning, graph problems including graph coloring, graph
matching, floor-plan design, temporal reasoning, spatial and
temporal planning, assignment and mapping problems, resource
allocation in AI, combined planning and scheduling, arc and path
consistency, general matching problems, belief maintenance,
experiment planning, satisfiability and Boolean/mixed equation
solving, machine design and manufacturing, diagnostic reasoning,
qualitative and symbolic reasoning, decision support,
computational linguistics, hardware design and verification,
configuration, real-time systems, and robot planning,
implementation of non-conflicting sensor systems, man-robot and
robot-robot communication systems and protocols, contingencytolerant motion control, multi-robot motion planning, multi-robot
task planning and scheduling, coordination of a group of robots,
and many others
• End of this part
Additional Slides
on Quantum
and emotional
robotics
• Huffman and Clowes
created an approach to
polyhedral scene
analysis, scenes with
opaque, trihedral solids,
next improved
significantly by Waltz
• Popularized the
concept of constraints
satisfaction and its use
in problem solving,
especially image
interpretation.
• Objects in this
approach had always
three plane surfaces
intersecting in every
vertex.
Constraint
Satisfaction Image
Analysis by Waltz
Constraint Satisfaction Image Analysis by Waltz
• There are only four ways to label a line in this blocks world model.
• The line can be convex, concave, a boundary line facing up and a
boundary line facing down (left, or right).
• The direction of the boundary line depends on the side of the line
corresponding to the face of the causing it object.
• Waltz created a famous algorithm which for this world model which
always finds the unique correct labeling if a figure is correct.
AC-3: State 2
3. Queue:
(2,3)(3,2)(3,4)(4,3)(4,
1)(1,4) (1,3)(3,1)
4. Removing (2,3).
5. L3 on 2 inconsistent
with 3, so it is
removed.
6. Of arcs (k,2), (1,2) is
not on queue, so it is
added.
Examples of CSP in robotics
• Scene recognition
• Motion generation in presence
of constraints
– internal (low power, don’t hit
itself)
– external (shape of racing track,
wolf-man-cabbage-goat)
• Gesture under emotions
• Communication in a swarm of
robots (graph coloring)
• Robot guard (set covering)
Adiabatic
Quantum
Computing
to solve
Constraint
Satisfaction
Problems
efficiently
Adiabatic Quantum Computing to solve Constraint
Satisfaction Problem efficiently.
• Will February 13th 2007 be remembered in annals of computing.?
• DWAVE company demonstrated their Orion quantum computing system
in Computer History Museum in Mountain View, California.
• The first time in history a commercial quantum computer was presented.
•
The Orion system is a
hardware accelerator
designed to solve in
principle a particular NPcomplete problem called
the two-dimensional Ising
model in a magnetic field
(for instance quadratic
programming).
•
It is built around a 16-qubit
superconducting adiabatic
quantum computer (AQC)
processor.
Orion computer from DWAVE
• Conventional front end
• The solution of an NPcomplete problem:
• 1. Pattern matching applied
to searching databases of
molecules.
• 2. Planning/scheduling
application for assigning
people to seats subject to
constraints.
• 3. Sudoku
7
1
3
2
5
8
2
2
9
8
3
9
8
3
7
4
5
7
4
5
6
6
4
6
2
Orion Is the Constraint Satisfaction
Solver
Does it have
quadratic speedup?
• The company
promises to
provide free
access by Internet
to one of their
systems to those
researchers who
want to develop
their own
applications.
Orion computer from DWAVE
• The plans are that by the end of year 2008 the Orion systems
will be scaled to more than 1000 qubits.
•
Company plans to build in 2009 processors specifically
designed for quantum simulation, which represents a big
commercial opportunity.
• These problems include:
protein folding, drug
design and many other
in chemistry, biology and
material science.
• Thus the company
claims to dominate
enormous markets of
NP-complete problems
and quantum simulation.
We plan to concentrate
on robotic applications of
the Constraint
Satisfaction Model.
• Adiabatic Quantum Computing was proved equivalent to standard QC
circuit model.
• Each of the developed by us methods can be transformed to an
adiabatic quantum program and run on Orion.
• We developed logic minimization methods to reduce the graph that is
created in AQC to program problems such as Maximum Clique or SAT.
• This programming is like on “assembly level” but with time more
efficient methods will be developed in our group.
– This is also similar to programming current Field-Programmable Gate
Arrays.
Future work on Adiabatic Quantum
Controller for a robot
• In the second research/development
direction the interface to Orion system
will be learned
• How to formulate front-end formulations
for various robotic problems as
constraint-satisfaction problems for this
system?
New Research Direction
• New approach to
quantum robotics based
on reduction to
Constraint Satisfaction
Model
•Well-known
problems
•New problems
Quantum
Emotional
Robots
Emotional
Robot Helpers
• Because humans attribute emotions to other humans and
to animals, future emotional robots should perhaps be
visually similar to humans or animals,
– otherwise their users would be not able to understand robots’
emotions and correctly communicate with them.
• Observe that the whole idea of emotional robot helpers is
to enable easy communication between humans and
robots.
Robot emotions
The research on robot emotions
and methods to allow
humanoid robots to acquire
complex motor skills is
recently advancing at a very
fast pace.
• Simple emotions like “fear” or “anger” or
behaviors like obstacle-avoidance for
wheeled mobile robots.
• Subsumption architecture.
• Practically insufficient to cover all
necessary behaviors of future household
“helper robots”.
• Larger biped robots are very
expensive
– hundreds thousands dollars.
• Recent small humanoid robots.
• We acquired two KHR-1
robots and integrated them to
our robot theatre system with
its various capabilities such as:
– sensors,
– vision,
– speech recognition and
synthesis
– Common Robot Language.
Emotions can be
best expressed
by a biped robot with
human-like face
• Humanoid
robots to
express
emotions:
• M. Lukac
uses humanlike faces and
head/neck
body
combinations.
• KAIST theatre
used wholebody
stationary
robots with
hands.
• Walking biped
robot can
express the
fullness of
human emotions:
–
–
–
–
body gestures,
dancing,
jumping,
gesticulating
with hands.
• Emotions can
be:
– Emergent Arushi
– Programmed –
Martin Lukac
ISMVL
– Mimicked –
ULSI
– Learned –
Martin Lukac
Reed-Muller
Synthesis of quantum circuits and
state machines from examples
– Quantum mappings –
Quantum Braitenberg
Vehicles – Arushi ISMVL
2007
– Quantum Oracles such as
Grover – Yale ISMVL 2007
– Emotional State Machines
– Lukac ISMVL 2007
– Quantum Automata and
Cellular Quantum
Automata – Lukac ULSI
2007
– Motion – Quay and Scott
Quantum Emotional Facial
Gestures
First View: Emotion as
synthesized behavior
Serchuk et al (2006) discuss emotion as mapping
from internal state to observable output behavior.
We want to design these mappings well, so that they
wil be similar to humans
Emotional state =
state of all emotion variables
Physical variables = positions,
speeds, accelerations, words,
Wheel of emotions
Active - Passive
Positive Negative
Internal representation
of emotions by vectors
in multi-dimensional
space
Mapping from internal
to external
representation of
emotions
Second View: Emotion as
emergent, evolvable behavior
• Here emotion is an emergent behavior that arises from sensors,
drives, effectors and logic.
• This may look like human, animal behavior but also as an entirely
new “other world” behavior, behavior as it may be.
Sensors, vision and fusion =
features and patterns
Drives and effectors
Main input-output mapping
(perception, internal state, behavior)
Precise motion generation (behavior)
Evolved “emotional”
behavior of robot
Human Emotions Perceived
by Robot
• Robot perceives emotions of a human
• Emotional aspect of speech
• Text from speech recognition
• (I hate you example)
• Facial gestures
• Body language and hand upper body gestures.
• Camera with software
• Microphones with speech recognition/speech
analysis system
You do not need robot, this
may be done by laptop with
microphone and camera.
Robot perceives human mood
• From top to bottom, the
continua
shown in each row are…
– happiness (H) - surprise (U)
– surprise (U) - fear (F)
– fear (F) - sadness (S)
– sadness (S) - disgust (D)
– disgust (D) - anger (A)
– and anger (A) - happiness
(H)
Why we need Robot-Generated
Emotions?
• Robot presents its emotions to a human
• Why we need it?
• Robot who helps elderly
• Assistive robot for disabled
• Robot that works with mentally challenged children (autism, Asperger
Syndrome, ADD),
• Robot receptionist
• Robot barman
• Robot astronaut helper
• Robot museum guide
• Robot theatre (mostly in interactive theatres)
•
•
•
•
Imitation of human emotions
Interaction with human based on emotion
Improvisation of theatrical plays, texts, stories
Interpretation of human behavior in psychological terms, negotiation and
cheating
Emotions in Humanoid
Robots
• Humanoid Robotics focuses on communication with
humans that includes:
–
–
–
–
Behavioral changes and emotional expressions,
Emotional alterations of text-to-speech,
Facial mimics and gestures,
Overall body language (posture) and hand upper body gestures
(hands, neck).
– Member postures and movements
66
Symmetry of emotion transmission
Robot
reconstructs
Human
emotion
Human
emotion
Human
behavior
Human
perceives
robot
behavior
and
emotion
Two aspects – two
approaches
Robot
perceives
Human
behavior
Robot
creates
its
emotion
Robot
expresses
its
behavior
Emotions as
emergent
behaviors
Emotions as
learned
behaviors
Traditional and modern
theories of emotion
• Observable (traditional) emotions: emotional
behaviors, moods, content changes (speech
variations, etc)
• Modern Hypothesis: emotions and feelings
are influencing decision making, problem
solving, memory efficiency and so on.
68
Two level representation of the
Cognitive-Emotional robot Structure
Flow of
emotions
Flow of
actions
Quantum
Mechanics
to model
emotions
Quantum Mechanics to
Model Emotions
• The problem being considered here is the synthesis of logic controller
allowing the robot to modify its actions and express unique emotional states
• The emotional expression is desired to be compelling the human user to
communicate with the robot,
– the behaviors should be original and non-repeating
• Standard classical approaches can be compared to an FSM approach;
– the robot action space (behavioral space) is a finite set of states that the robot
learns or just uses in a input driven mapping
71
Concept: Emotional Quantum
State Machine
• Design a machine that will simulate the
articulation of human social behavior:
– Subjective
– Non repetitive
– Innovative
• But still:
– Socially acceptable or not
– Behaviorally understandable
– Safe (the framework of this behavior is
purely virtual – no contact)
72
Quantum Hierarchical Model of Emotions
• Because the concept of emotional expression can
be extended to a functional model; emotional
expression affects the robot functioning.
• Here the concept of QFSM is extended to a
Quantum Cellular Automata based on the quantum
emotional state machine
• The quantum string rewriting is extended to a
complete robot hierarchy rewriting schema
73
One Machine Model in Our
Approach
Q  quantum states
S  classical states
E  alphabet
Θ  quantum evolution function
δ  state transitio n function
q0  initial quantum state
s0  initial classical state
S acc  set of accepting states S_acc in S
S rej  set of rejecting states S_rej in S
M = Q, S, E,Θ,δ, q0, s0, S acc , S rej 
Definition of Emotion
• Emotion is the result of measurements of a hybrid
classical/quantum system
• In terms of quantum mechanics, emotions are
represented by quantum states that we (observe)
know only after measurement,
– but we can operate on them deterministically in the
Hilbert space.
• Emotional evolution is represented by quantum
operator (unitary and non-unitary, including the
measurement)
Simulating emotions for practical
applications
• Simulating emotions as only observable behaviors is not sufficient to
make emotional robots
• Definition: Emotional State Machine is a model of FSM that can modify
its state and output independently of the content of the input, but based
solely on its current state.
• Definition: Robotic Emotions are simulated emotional states allowing
the robot to perform a given action in a way that satisfies its current
emotional state.
• We propose a emotional model as computational process distributed
across the robot software controller allowing to use emotions to modify
all robot actions
76
Emotional Quantum Automata
•
•
A network of Emotional State Machines is called an Emotional Quantum Automata
(plural - per similarity to Cellular Automata)
This network can be regular or not.
•
If regular, it can be:
–
–
–
–
•
One-dimensional and one-directional (pipelined)
One-dimensional (like one-dimensional cellular automata)
Two-dimensional (like Game of Life and Two-Dimensional cellular Automata of Wolfram)
more dimensional.
Emotional Quantum Automaton is therefore a generalization of Cellular Automata,
Random Boolean Networks and Quantum Cellular Automata
Example of onedimensional and
one-directional
Cellular
Quantum
Automata
Complex space vs. Real States
Neighbors
Emotional states
travels across the
robot body
Time
Simulation of a quantum emotional automata. The dots represents real states (observables)
and the surrounding represents complex components.
Of interest is the fact that even in a variants that the automata communicate exclusively via
classical data channels, the complex part can travel through space.
Thus, emotions encoded by such a complex state can be moving across robot controller and
create unexpected local effects
78
Emotional Parameters for control
in Cynthea Robot
Device
Parameters
Other
This slide shows which parameters are affected
by which input and output devices
Emotional Model
Emotional
Robot controller
Robot controller
Emotional
Robot sensors
Robot sensors
Emotional
Robot actuators
Robot actuators
Formal Language
• Each element in the robot is represented as Quantum
Emotional State Machine (EQSM), such that on each
level of robot control hierarchy the emotions can
influence both the visible (perceptible) and the nonvisible robot processes
80
Emotional Model
• Energy – simulated energy representing the emotional state
of the robot
Emotional State
Energy
• Strategy – the translation function mapping the emotional
state to a state parameters, (function dependent)
Emotional State
State
Parameters
• The input command - represents the robot command such as one
obtained from a sensor (user input, other robot), specified in the
CRL language
• The emotional parameters translated to particular variables, are
used to modify the global state of the robot and also the local
function (Command rewriting)
Emotional
Parameters
State of the robot
81
Project
Overview
• KHR-1
– Biped robot
– 17 servos
– 2 RCB-1 servo
controllers (each
12 servos)
– Serial port
connectivity
Common Robot Language.
• We developed symbolic approach
to robot specification based on a
Common Robot Language.
• While the syntax of this language
specifies rules for generating
sentences, the semantic aspects
describe structures for
interpretation.
• Every movement is described on
many levels, for instance every
joint angle or face muscle are at
low level and complete movements
such as pushups or joyful hand
waving are at a high level.
Common Robot Language.
• These aspects serve to describe
interaction with environment at
various levels of description.
• It uses also the constraint
satisfaction problem creating
movements that specify constraints
of time, space, motion style and
emotional expression.
Describing movements, behaviors
and emotions
• The goal of our Common Robot Language is to describe
human-oriented movements
• But it exceeds these behaviors to those like
anthropomorphic animals and fairy tale characters.
• We created new GUI interface and robot controlling
language specific to KHR-1.
– Editing functions.
– Testing functions.
– The ability to read information back from the robot by serial
communication was added.
• There are two main functions that we achieved:
– mimicking,
– behavior state machine.
Using HBP robot vision software for
human mimicking.
• Control behaviors mimicked from a human standing in front of
the camera.
– (with state machine or not)
• We wanted the KHR-1 to mimic human motion that was being
shown on the screen by the HBP software.
• The HPB works by taking an image of a person’s upper body. It
then will try and identify the face.
• Once it can recognize a face it will then look at the body.
• The image that it acquires is converted to a set of feature
(parameters) values assigned to several groups of variables.
What is wrong with our vision
software?
•
•
•
•
HBP is slow
OPENCV is slow
Robot responds with delay
HBP is not accurate
•
That one great thing about HPB, is that you have the
option of modifying the original code to some extent and
make your own features.
• To speed up the image recognition we will use the
Orion quantum computer in the next project
Genetic Algorithm for Motion:
Chromosome
• The chromosome is an array of 0’s, 1’s, and -1’s.
These values control the 12 servos that dictate the
movement of the robot. ( -1 = full left, 1 = full right, 0 =
middle position).
• Each chromosome holds 4 sets of servo commands to
the 12 Servos. So, each chromosome is 48 bits long.
• These 4 sets of commands to the servos will generate
a sequence of movements that forms a gesture.
Genetic Algorithm:
Chromosome
• Below is a visual representation of the 48 bit chromosome
• Each rows each represent a command to the 12 servos. The 4 rows
together will send 4 commands to the 12 servos that will form a
gesture
• Orange colors sections = Right arm control bits, Blue color
sections = Hand control bits, Green color sections = Head control
bits
Research
Topics
Research areas
1. Quantum Braitenberg Robots
2. Quantum Subsumption Architecture
3. Quantum Emotional Robotics
4. Other quantum robot architectures such as probabilistic
5. Quantum Search
6. Quantum Image Processing and Pattern Recognition
7. Quantum Spectral Transforms
8. Quantum Games
9. Quantum Theorem Proving
10. Quantum Learning (QNN, Quantum Automata, Markov Models,
Bayesian Networks, Associative Memories, etc.
11. Quantum Holographic memories
12. Models of Quantum Physical Processes in biology, chemistry,
etc.
Disclaimer – do not worry!
• We talk here about emotions of these:
And not these
Words such as
“memory, emotion,
knowledge,
remember, solve,
prove” carry humanlike meaning but with
time we are used to
use them in a broader
sense
Our Main New Idea:
Two Layer Action-Emotion FSM Model
Emotions are not
something “additional”
to rational thinking and
acting
Emotions are
intimately intertwined
in every process of a
robot on any level of
hierarchy
Simplified model of
Emotional State
Machine
Instead of a hierarchy
of state machines we
have a hierarchy of
Emotional State
Machines
Emotional State Machine
Deterministic classical
physics/compute science world
(Turing compatible)
Measurement of
the machine state
Quantum
memory
Emotional
evolution
(operator)
Quantum (Hilbert Space)
(Quantum Turing Machine
Compatible)
Quantum computing Basics
In quantum computing the system (circuit) is represented in the form
of a wave: | ψ >=  ci | i > , |ci |2= 1
i
The space of the system is complex Hilbert space H of dimension N.
The basis states are orthonormal, for boolean logic:
1
 0
| 0 >=   , | 1 >=  
 0
1
| ψ >= α | 0 > +β | 1 >
The operations on the system are in the form of Unitary matrices
being rotations of the state vectors in the space H: U | ψ >=  ci | mi >
i
To retrieve the result the system has to be observed or measured.
Measurement is an outside operation on the system, and destroys the quantum state.
This operation projects the system onto real basis states such as defined above.
Because the measurement is completely random, the information is extracted from the
collapsed state that has the form of:
M |ψ >
| ψ' >=
95
< ψ | M *M | ψ >
Quantum Computing Basics (contd.)
Special phenomena can be observed in quantum:
● The system can be in superposition (being in all states at the same
time) | ψ >=  c | i >, |c | = 1
● The system can be entangled, the outcome of the whole system or
of its subparts is dependent on the measured output qubit(s).
2
i
i
i
M 0 = 1

0

0

0
1

1 0
M 0 | φ >=
2 0

0

0 0 0

0 0 0

0 1 0

0 0 0
0
0
0
0
0
0
1
0
| 00 > + | 11 >
| φ >=
2
0  1 
1
 
 
0  0  1  0 
=



0  0 
0
2
 
 



0
0  1 
 
1
 
1 0
2 0
 
0
 
2
1
 
 0
= 
0
 
1
 
Despite the fact that before
measurement both qubits
have the probability of 0.5 of
being in state 0 or 1, after one
of the qubit is measured the
state of the second one is
instantaneously determined
96
Disclaimer: Definition of Emotion
• We use, among other concepts, the quantum
concepts to define and use emotions
• In our model emotions are formally defined, you
can think about them as quantum states or
quantum operators.
• Then, in this work there is no implication that our
“emotions” are related to human emotions
– other than that we want to emulate human behavior by a
humanoid robot.
So what are robotic emotions?
Motivations for Emotional Robotics
• Human-Human interaction is highly variable,
individual, unique, non-repeating, etc.
• Emotional Robot, Humanoid Robot
• Quantum emotional state machine
• Control logic for robotic quantum controllers in
order to increase interactivity and quality of
communication
• Logic synthesis of such circuits is in the middle
of this paper
Emotion
al Robot
Emotion
Recognition
versus
Emotion
Generation
• This generic situation, where
the robot’s behavior is
conditioned upon the input from
the feature detectors connected
to the camera, maps to a
constraint satisfaction problem as
described here.
• The way this would work is that
the human / camera / robot
system would generate
optimization and satisfiability
problems, to determine how the
robot’s effectors should fire, and
these problems can be remotely
solved using Orion.
• For example, you could acquire
a Hansen Robotics Einstein, sit it
him on your desk, train a camera
on your face, use an anger
feature detector that causes the
Einstein robot to laugh harder the
angrier you get.
Moral
Robot
Concepts of a moral robot
1. So far, moral robots are
built for military (immoral
robots?)
2. Our concept is a moral
robot , helper,
housekeeper and
entertainer for an old
demanding lady.
1.
2.
3.
Path planning and tasks
execution
Access to internet
Conversation and
dancing
1. Robot integrates modal
and deontic logic.
2. Principles of morality:
1.
2.
3.
4.
Be nice
Help and protect
Never harm by accident
or mistake
Resolve conflicts of the
lady with family and
friends
3. Asimov’s Laws.
4. Exclusions and special
cases.
Multiple-Valued Logic
in Quantum Domain
Oracle for Quantum Map of Europe Coloring
A B
0
1
2
3
0
1
2
3
0
1
2
3
1
0
3
2
2
3
0
1
3
2
1
0
A B
0
1
2
3
0
0+1=1 1+1=0
2+1=3
3=1=2
1
0+0=0
1+0=1
2+0=2
3+0=3
2
0+3=3
1+3=2
2+3=1 3+3=0
3
0+2=2
1+2=3
2+2=0
3+2=1
A
A
+1
B
0
+1
1
0
B
1
+2
2
+3
2
+3
3
+2
3
Quaternary Feynman
1 when
A=B
Quaternary input/binary output
comparator of equality
Oracle for Quantum Map of Europe Coloring
Comparator for
each frontier
A
+1
B
0
1  1 -- when control 1
1
+3
2
1 -- for controls 0,2 and 3
3
+2
1
1
+1
1
1
+1
Binary qudit =1 for
frontier AB when
countries A and B
have different colors
C
+1
D
0
1
+3
2
3
+2
0
Quaternary
controlled binary
target gate
Binary Toffoli
Binary signal
1 when all
frontiers well
colored
Conclusions and future work.
Didactic Aspects
•
KHR-1 is now able to mimic upper body human motions.
• Students who work on this project learn about robot kinematics, robot
vision, state machines (deterministic, non-deterministic, probabilistic
and quantum - entangled) robot software programming and
commercial robot movement editors.
• The most important lesson learned is the integration of a non-trivial
large system and the appreciation of what is a real-time
programming.
• It is important that the students learn to develop a “trial and error”
attitude and also how to survive using a non-perfect and incomplete
documentation.
• It was also emphasized by the professor that students create a very
good documentation of their work for the next students to use.
Conclusions
1. Our goals are to both create a model innovative
robot theatre and a theory of robot theatre that
would be similar to the theory of film or theory of
interactive computer games.
2. We believe that robot theatre will become a new
art form and we are interested what are the basic
questions related to the art of performing robots.
3. We hope to have an interesting feedback to our
ideas from all groups of PSU researchers.
New Research Area or only
application?
1.
2.
3.
4.
What is Robot Theatre Theory?
What are its main methods?
How to evaluate Robot Theatres?
Is robot theatre only an application of
robotics or is it more? What ?
•
We see that humans can laugh looking at our
theatre.
Will we ever experience humans crying at robot
performances?
What can we learn from robot theatre that is
not a standard robotics problem?
•
•
1.
Final Fundamental
Questions:
Will Robot
theatre be ever as popular
art form as film or theatre?
2. Will robots be popularly used in
theatres?
3. Will we see robot theatre toys?
4. Will home robots become also robot
entertainers?
Constraint Satisfaction for Robotics
• Insufficient speed of robot image processing and pattern
recognition.
– This can be solved by special processors, DSP processors, FPGA
architectures and parallel computing.
• Prolog allows to write CSP programs very quickly.
• An interesting approach is to formulate many problems using the
same general model.
• This model may be predicate calculus, Satisfiability, Artificial
Neural Nets or Constraints Satisfaction Model.
– Constraints to be satisfied (complex formulas in general)
– Energies to be minimized (complex formulas)
SAT as a constraint satisfaction
problem
(a + b’ + c) * (b + d’) … = 1
=1
=1
Highly parallel
system of
updating nodes
c
a
=0
=1
b
d
(a + b’ + c)
(b + d’)
Yes , do nothing
No , update
to nodes
nodes
SAT as a constraint satisfaction problem
(a + b’ + c) * (b + d’) … = 1
Constraints:
(a + b’ + c) = 1
(b + d’) = 1
…..
Energy optimization:
(a + b’ + c) = f1
(b + d’) = f2
….
Min ( f1’ + f2’ + …..)
Orion
programming
is just writing
equations for
constraints
and
equations for
energy
Conditional robot response based
on camera input
• This one is really cool.
• At a high level the way it works is as follows.
• You have a camera trained on a human. The data taken by
the camera is processed so as to detect features, which are
generalized patterns of behavior.
• For example, a feature detector could be configured to
detect the presence of anger in the human, for example by
learning-based methods.
• In addition, there is a robot, which is connected to the data
processing system connected to the camera.
• This robot has effectors which control its actions.
• In this application, the effector controls are functions of the
processed input from the camera, where the rules
connecting the two are user-determined.
High Level
Schematic
of DIM robot
Many research and teaching areas
My robotics classes and topics of theses are much broader than in other
universities:
1. Classical Robotics
2. Speech synthesis and analysis
3. Robot vision
1.
2.
4.
5.
6.
7.
Motion generation
Dialog and natural language
Scripts and agents.
Search and Machine Learning.
1.
2.
3.
4.
8.
Face recognition and tracking,
gesture recognition, etc.
Search theory
Decision trees and constructive induction (AC decomposition)
CSP
Pattern Recognition
New models:
1.
2.
Probabilistic robotics,
Quantum Robotics,
Many partial practical projects,
broad range of technical
knowledge
1.
2.
3.
4.
5.
6.
7.
8.
9.
Controllers and processors
FPGA and PLD, VHDL and Verilog.
CUDA and GPU
Motors and sensors
RobotC, C, C++, Prolog, LISP, PYTHON, etc.
Writing requirements
Creativity and invention.
Integration of engineering, science and art.
Many awards for high school students.
Conclusions
1. Our goals are to both create a model
innovative robot theatre and a theory of
robot theatre that would be similar to the
theory of film or theory of interactive
computer games.
2. We believe that robot theatre will become a
new art form and we are interested what are
the basic questions related to the art of
performing robots.
3. We hope to have an interesting feedback to
our ideas from the System Science oriented
Questions:
1. Will Robot theatre be ever as
popular art form as film or theatre?
2. Will robots be popularly used in
theatres?
3. Will we see robot theatre toys?
4. Will home robots be also
entertainers?