* Proposal for PhD degree in Electrical & Electronic Engineering

By research

Title: Intelligent Robots Using Artificial Intelligent

(Multi-Agent robots using Artificial Intelligence)

Submitted to
Sudan University of Since and Technology
Faculty of Engineering
Electronics Department

By:
Mousab Mohammed Hassan Mahgoub
-Introduction
Introduction
-Problem Definition
-Literature Review and Relevant
Studies
-Objectives
-Methodology
-Expected Outcome
-Study schedule
-References
 Introduction
Artificial intelligence (AI) and increasingly complex algorithms
influence our lives and our civilization more than ever. The areas of AI
application are diverse and the possibilities far reaching: in particular,
because of improvements in computer hardware, certain AI
algorithms already surpass the capacities of human experts today.[1]
Research in AI has built upon the tools and techniques of many
different disciplines, including formal logic, probability theory, decision
theory, management science, linguistics and philosophy. However,
the application of these disciplines in AI has necessitated the
development of many enhancements and extensions. Among the
most powerful of these are the methods of computational logic. I will
argue that computational logic, embedded in an agent cycle,
combines and improves upon both traditional logic and classical
decision theory. I will also argue that many of its methods can be
used, not only in AI, but also in ordinary life, to help people improve
their own human intelligence without the assistance of computers.[2]
Computational logic, like other kinds of logic, comes in many forms. I
will focus on the abductive logic programming (ALP) form of
computational logic. ALP in an agent cycle is a powerful model of
both descriptive and normative thinking. As a descriptive model, it
includes production systems as a special case; and as a normative
model, it includes classical logic and is compatible with classical
decision theory. These descriptive and normative properties of the
ALP agent model make it a dual process theory, which combines both
intuitive and deliberative thinking.[3]
Many of the multi-agent coordination issues such as task allocation,
path and trajectory planning, formation optimisation and pursuit-
evasion strategies are now becoming well understood, even though
demonstration of them using real robots in outdoor and unstructured
environments (i.e. as opposed to using simulation) are rare. More
recent research has focused on motion coordination within the
context of behaviour coordination such as target search and feature-
tracking behaviours. As a result, research into path planning and
control, multi-robot task and resource allocation, behaviour
coordination and communications has become coupled.
This is largely because the structure of the multi-robot cooperative
changes with time and the properties of the cooperative change with
structure. That is, the position of a robot within its team’s structure
and relative to its goals determines its projected sensing options,
prospects for information gain, and capacity to (say) accurately
explore, map and locate key features in its environment; just as its
inherent capabilities, sensing options, scheduling of sensors, and so
on impact the potential agent trajectories, behaviours, feature
tracking accuracy, communications strategies, etc. [4]
 Problem Definition
* The criterion for intelligence is related to statistical and economic
notions of rationality colloquially, the ability to make good decisions,
plans, or inferences. The adoption of probabilistic representations
and statistical learning methods has led to a large degree of
integration and cross-fertilization between AI, machine learning,
statistics, control theory, neuroscience, and other fields such as
speech recognition, image classification, autonomous vehicles,
machine translation, legged locomotion, and question-answering
systems.[7]
* AI researchers have worked on methods for solving several types of
combinatorial problems. Their efforts have been directed at making
the time-versus-problem-size curve grow as slowly as possible, even
when it must grow exponentially. Fuzzy logic theory is used to
estimate the system dynamic. To estimate the system dynamics
and eliminate the chattering, fuzzy variable structure controller is
commencing. This methodology is based on applied fuzzy logic and
artificial intelligent in equivalent nonlinear dynamic part to estimate
unknown parameters.[5]
* A major bottleneck in the design of multiagent systems is the design
of the protocols of interaction among their member agents.
Unfortunately, while individual agents are usually described in terms
of their knowledge, intentions, and know- how, extant approaches to
understanding the interactions among them are not able to make
full use of those abstractions. Even fairly recent research, which
provides primitives for communication among agents, has tended
to be con- cerned with the workings of the TCP/IP and similar
protocols.
* This lack of a general theory of the interactions among agents
forces the system designer to think in terms of what are, from the
point of view of muitiagent systems, merely details of the underlying
architecture. These details are important, but are simply out of place
at the level at which we wish to study such systems, The semantics
of these interactions is embedded in different procedures, some of
which involve network and operating system code. This makes the
nontrivial task of validating and modifying multiagent systems even
more difficult.
 Literature Review and Relevant Studies
• I have sketched two ways in which the ALP agent model, building
upon many different developments in Artificial Intelligence, can be
used by ordinary people to improve their own human intelligence. It
can help them express their thoughts more clearly and coherently, and
it can help them make better choices. I believe that the application of
such techniques is a fruitful direction of research for the future, and a
promising area for collaboration between researchers in AI and
researchers in more humanistic disciplines.[6]
• In this paper, we put forward the environment as an explicit part of
MAS. Furthermore, we defined environment as a first-class
abstraction in MAS that has a dual role. First, the environment is an
essential part of every MAS. The environment is the part of the world
with which the agents interact, in which the effects of the agents will
be observed and evaluated. It is the glue that connects agents into a
working system. Second, the environment provides a design space
that can be exploited by the designer. Distinguishing between agent
and environment responsibilities supports separation of concerns in
MAS, and helps to manage the enormous difficulties involved in
engineering complex real-world MAS applications.
 We have identified three levels of support that can be provided by
the environment: (1) direct access to the deployment context, (2) an
abstraction level that shields agents from the low-level details of
deployment, and (3) mediated interaction between agents.
Important responsibilities of the environment include:(1) the
environment provides structure for the MAS as a whole; (2) the
environment embeds resources and services; (3) the environment
can be responsible for maintaining dynamics in the system that
happen independent of agents; (4) contrary to agents, the
environment tis observable;(5) the environment is locally accessible
to agents; and finally (6) the environment can define different types of
rules on all the entities in the MAS.
The environment offers opportunities for all types of agency, and as
such provides a challenging area for synergetic research in the
interest of MAS in general. Considering the environment as a first-
class abstraction in MAS opens up new horizons for research and
development in MAS[7].
It has been argued that very general and capable AI systems
operating autonomously to accomplish some task will often be
subject to effects that increase the difficulty of maintaining
meaningful human control (Omohundro 2007; Bostrom 2012, 2014;
Shanahan 2015). Research on systems that are not subject to these
effects, minimize their impact, or allow for reliable human control
could be valuable in preventing undesired consequences, as could
work on reliable and secure test beds for AI systems at a variety of
capability levels.[5]
This research will address these coupled tasks within a single
framework, whereas previously the problems of and approaches to
communications and sensor scheduling, feature tracking and
trajectory control were largely de-coupled and addressed using
independent algorithms and strategies and then combined using
some form of executive controller or architecture. More recent work
attempts to manipulate the sensing process in order to maximise the
information gain and feature location estimation, without using any a
priori information. When the sensors are passive, this introduces a
number of aspects that are not under sensor control (i.e. when
precisely observations are made and what the observations are of),
both of which have an impact on the development of longer-term
scheduling strategies for the sensors and agent. This research will
coordinate groups of agents into teams, the members of which will
take (or will be instructed to take) a decision to commit to a particular
task but who will receive common reward for task achievement as a
result of team decisions.[8]
* Objectives:
 
• To identify an application of machine intelligence and robotics by
design Multi agent mobile robots that can be remotely controlled
from a long distance.
• Design an autonomously navigator robots with Deductive learning
to do difficult task.
• The robots can coordination and cooperative to avoidance obstacle
in dynamic environment.
• To estimate the benefits of successful adoption of machine
intelligence and robotics techniques and to prepare forecasts of
their growth potential.
• To recommend program options for research advanced
development and implementation of machine intelligence and robot
technology for use in program planning.
 Methodology:

The first significant application of artificial intelligence is machine

learning. The main approach for machine learning is deductive
learning. Initially, the basic technique of deductive learning is rote
learning. As Firebaugh states, the basic level of learning known as
rote learning for both human and machine learning. In machine
learning it requires that knowledge is shown without being changed in
the system and no further process is desired , Inductive learning is the
other important approach for machine learning. For instance, active
learning is an advanced learning system that is effectively applied as
inductive learning. Settles stated that, a machine learning algorithm
that works with higher efficiency with less guidance labels when it is
free to select the data from the source is the main concept of active
learning.
Active learning is an impressive and beneficial technique that can be used
in numerous scenarios with less effort. Furthermore, rule-based learning is
another powerful inductive learning process. Rule- based learning
algorithms are well-performing and easy to compact machine learning
algorithms that approximate the values of a real valued functions from
multiple input variables by using ordered samples of decision rules. [10]
The rules:
• only one event happens at a time, which entails that only one agent acts
at a time;
• events have precisely determined effects; and events are necessarily
associated with a state change
A rigorous technical framework that
- Allows concurrent actions by multiple agents
- Allows actions to be of varying durations
- Allows nondiscrete models of time, which underlies actions
- Carefully relates actions to the temporal aspects of the framework
- Admits a notion of weak determinism in which different possible choices
of agents can be simultaneously captured
- Defines abstract actions or strategies such that they can coexist with
basic or primitive actions
 
* Expected Outcome
Design a robust artificial intelligent nonlinear controller for second
order nonlinear uncertain dynamical systems is one of the most
important challenging works. This project focuses on the design of a
robust chattering free mathematical model-base artificial intelligence
(fuzzy inference system) variable structure controller (MFVSC) for
highly nonlinear dynamic continuum robot manipulator, in presence of
uncertainties. In order to provide high performance nonlinear
methodology, variable structure controller is selected. This research
is used the artificial intelligence (fuzzy logic) theory. The results
demonstrate that the model base fuzzy variable structure controller
with switching function is a model-based controllers which works
well in certain and partly uncertain system.[11]
Study schedule:
 
Expected Outcome Year 1 Year 2
 
 
                 
Literature Review and Constraint
               
formulation
                 
Provisioning of protocol and assumptions
               
 
                 
Simulator design and code verification
 
                 
Simulation evaluation and test
 
                 
Design and implementation Robots
 
                 
Results analysis
 
                 
Thesis writing
 
 References
 
[1] Artificial Intelligence: Opportunities and Risks, Policy paper,
Foundational Research institute.
[2] Robert Kowalski, Imperial College London United Kingdom,
Artificial Intelligence and Human Thinking.
[3] Artificial Intelligence and Human Thinking, Robert Kowalski,
Imperial College London, United Kingdom.
[4] Sensor Coordination for Multi-Robot Teams, Prof Anthony Finn ,
Suitable as PhD
[5] Developing an Artificial Intelligence Engine,Michael van Lent and
John Laird ,Artificial Intelligence Lab, University of Michigan 1101
Beal Ave. Ann Arbor, 2110.
[6] Shifting viewpoints: Artificial intelligence and human–computer
interaction Terry Win grad Stanford University, 353 Serra Mall,
Stanford, CA, USA Available online 1 November 2006.
[7] Auton Agent Multi-Agent Syst (2007) , Environment as a first class
abstraction in multiagent systemsDanny Weyns · Andrea Omicini ·
James Odell
[8] [ Sensor Coordination for Multi-Robot Teams ,Prof Anthony Finn ,
Suitable as PhD 2015
[9] Research Priorities for Robust and Beneficial, Artificial Intelligence,
Stuart Russell, Daniel Dewey, Max Tegmark.
[10] Application of Artificial Intelligence in Robotics, Abubakar Sadiq
Usman Bello,2010.
[11] Shifting viewpoints: Artificial intelligence and human–computer
interaction Terry Win grad Stanford University, 353 Serra Mall,
Stanford, CA, USA Available online 1 November 2006.