Transcript Intelligent Robotic Sensor Agents for Environment Monitoring

Intelligent Robotic Sensor
Agents for Environment
Monitoring
Presented by:
Yousuf Ajmerwala
Intelligent Environments
Copyright 2003 © CSE - UT Arlington.
Overview
• What is an Intelligent Robotic Sensor ?
• Intelligent Sensor System
• Applications in an Intelligent
Environments
Definition in context of an
Intelligent Environment
• Robotic Intelligent Sensor Agent:
– Capable of storing data
– Capable of processing data
– Capable of motion
– Acting on the environment to improve the
experience of its users
System Architecture
• Provide task directed information
gathering
• Model based multi sensor fusion
• Agent based resource management
framework
• Redundant multi carrier communication
systems
System Architecture
• Intelligent task directed information
gathering
– Selective environment perception
• Focus on parameters important for a task
• Avoid irrelevant data
• Efficient use of each sensor
System Architecture
• Model based multi sensor fusion
– System that has sensors covering all aspects
of the geometric environment perception
process: far away, near to, touching, and
manipulation
– This model will provide modularity, plug-nplay and transparency for sensor fusion
and knowledge extraction
System Architecture
• Multi Sensor Fusion Framework
– Organize data collection from different types
of sensors
– Produce local and global world models
– Continuously update world models
By integrating heterogeneous sensor data into
a composite model of 3D object shape,
surface and material properties, such as
texture, elasticity, heat transfer
characteristics, radiation, and contact forces
System Architecture
• Agent based resource management
framework
– Division of a task into sub tasks
– Scheduling of the sub tasks
– Each sub task assigned to a specialized
R-ISA
System Architecture
• MPLS – Multiple Protocol Label Switching
– Integration of Layer 2(ATM) and Layer 3 (IP)
– Any selection of Layer 2 technologies and
Layer 3 protocols to provide a certain quality
of service
– MPLS – convenient for time sensitive sensors
Development & Evolution
• R-ISA’s should be able to cannibalize/
recycle other agents
– Modular Reconfigurable structures
• E.g., recycle dead agents
– Energy reserves
– Spare parts
• E.g., repair wounded/dead agents
– Reincarnation of a new agent
– Status advertising mechanism
• E.g., discover agent functionalities
• E.g., monitor health of other agents
Development & Evolution
• Cooperating Agents
– Able to work together with both Mobile &
Static agents to achieve a specific goal
• E.g., two R-ISA’s could assume best relative
positions to take the stereo image of an
environment
• E.g., one R-ISA could illuminate the subject
so the other could take the image
Development & Evolution
• Evolutionary Mechanism
– Allow transfer of learning and behavioral
genes (Knowledge Base) obtained via
experience from one R-ISA to another RISA.
Relevance to Intelligent
Environments
• Monitor environment parameters to
maintain a natural living environment
Conclusion
• Even though the field of robotics has
advanced, there is work to be done to
address the communication needs of the
agent
– Communication protocols
– Developing communication grammer
References
• Wireless Intelligent Sensor Network for
Environment Monitoring
• Petriu, Patry,Whalen, Daher, Groza.
Intelligent Robotic Sensor Agents for
Environment Monitoring
Q&A
• ??