Robot Autonomous Perception Model For Internet-Based Intelligent Robotic System

1
Robot Autonomous Perception Model For
Internet-Based Intelligent Robotic System

• By
• Sriram Sunnam

2
Introduction

• Internet based robot
• A platform build to provide information for robot
  can increase its intelligence and performance of
  teleoperation greatly

3
Agenda

• Problem definition
• Robot Autonomous Perception Model (RAPM) is
  proposed
• Optimizing paths based on Genetic Algorithm are
  discussed
• Experimentation part is presented
• Conclusion and future work

4
Mobile Robots

• Limitations of Mobile Robots
• Limitation of data handling capacity
• Cannot process excess data, decision and
  controlling
• Difficulty to fulfill complex work using only its
  own sensors

5
Solution

• System that allows a robot manipulator to be
  teleoperated via the world wide web.
• WWW provides low-cost and widely available
  interface that can make teleoperated resources
  accessible to a broad range of users.
• Intelligent Space- Navigation platform for
  human and robots where Distributed Intelligent
  Networked Devices (DIND) can detect position of
  human or robot by their color bars.
• Detected code is transformed into coordinate data
  and fed back to human or robot.

6
Problems in Internet based robots

• Random Delay (internet)
• Limited Bandwidth (Internet)
• Loss of data (Internet)
• Thousands of Administrators (Internet)
• Failures in Laboratory Systems (Ex. route
  failures, service break off, electricity failures)

7
Solution

• Adopt new method to monitor real-timely, the
  information of the dynamic network
• Then issue the information to the network user.
• Offer specific information service that
  establishes an information channel between the
  service and the network users
• Thus accomplish communication, interactive
  operation, controlling between robots and sensor
  and actuator.

8
Proposed Solutions

• Robot Autonomous Perception Model (RAPM)
• Searching for resource in the system, optimizing
  paths based on Genetic Algorithms

9
RAPM

• RAPM (Robot Autonomous Perception Model) uses
  JINI
• JINI is a middleware to build distributed system
  through java
• JINI Socket and RMI (Remote Method Invocation)
  communication protocol
• JINI can implement plug-and-play mechanism and so
  allows dynamic expansion ability

10
RAPM

• Lookup Service It confirms any new services
  being added to the network. Connections between
  the system and user should be realized through
  Lookup Service.
• Discovery RAPM entity ( ex video server)
  before using other services, must look for a
  valid lookup service from different RAPM colonies
  which is called discovery
• Lookup It is a catalogue service in a RAPM
  colony used to obtain new Lookup Service.
• Issue Admission of an equipment or application
  program into a network after its service finds
  the network
• Invoke The agent that is used by the network to
  communicate with the service.

11
RAPM
Jini Reggie Server
12
Optimizing Paths Based on GA

• Dynamic network can transfer into a temporary
  static network through the RAPM
• Several robots use same nodes (start and end) at
  same time.
• Paths optimized for all mobile robots so that
  robots use the same nodes at different time or
  use the different nodes.
• Use Genetic Algorithms to optimize the paths

13
Optimizing Paths based on GA

• Fitness function Used to avoid robots using the
  same nodes at the same time
• Sum of all nodes that robot will pass through
• Sum of all nodes that robot will pass through at
  the same time
• Influence between robots

14
Optimizing Paths based on GA
15
Experiment

• Pioneer2DX mobile robots of ActiveMedia Company
  used with Sony EVID31 camera as vision sensor.
• Aim Two robots search a blue goal and an yellow
  goal
• Cannot find the goals with their own sensors
• Vision sensors access the network or internet
  through computers
• These vision sensors and computers compose video
  servers

16
Experiment

• Jini server is started
• Two robots and video servers join Jini network
  and register information in the Jini server
• Robots get information of goals and video servers
• Optimize the paths based on GA.
• Robots move toward the goal

17

• Path of robot 1 A-E-F
• Path of robot 2 B-C-E
• When robots are close to goal, they use their own
  vision sensors to search, approach and finish the
  mission

18
Conclusions

• Proposes the Robot Autonomous Perception Model
  (RAPM).
• Internet based robot can perceive and utilize the
  information resource as video server.
• Optimizing the robot paths using GA can get
  better results.
• Future work includes including more devices and
  algorithms into the jini system to provide more
  information resource to robot to improve its
  intelligence

19
Comments

• In future internet could be used for dynamic
  identification of objects
• Internet improves the intelligence of the robots
  to very high level and so this intelligence could
  be used by the robots to do lot of things with
  their own intelligence

20

• Thank You

21

• Questions?