Sensor Networks Based on Biological Architectures

1
Sensor Networks Based onBiological Architectures
2
Scenarios
Atmospheric Research
Planetary Exploration
Industrial Monitoring
3
Assumptions

• Autonomous sensor network
• Self-deploying, self-monitoring, self-repairing
• Variety of functional devices
• Carriers, data collectors and transmitters,
  sensors
• Variety of sensor devices
• Position sensors, cameras, temperature sensors,
  spectrometers, etc.

4
Properties of Biological Architectures

• Self-Organization
• Fault-tolerance Survivability
• Evolution

5
Mechanisms of Biological Architectures

• Migration
• Birth/death Renewal
• Diversity in Reproduction
• Pheromone Communication
• Food

6
Example 21st Century Planetary Exploration
The Mars Microprobe Project will demonstrate key
technologies for 21st century "network" missions,
in which multiple microprobes or landers,
released from a single spacecraft, will provide
comprehensive studies of dynamic, complex
phenomena, such as climate systems and seismic
activity. -NASA
7
Goals

• Geographical mapping
• Exploration and mapping of topology, surface
  conditions, density, etc.
• Distributed sample/data collection
• Collecting soil samples, images, atmospheric
  measurements, etc.
• Finding target object types
• Detecting water, minerals, geothermal activity,
  life, etc.
• Performing experiments

8
Types of Sensor Agents

• Oribiting station
• Satellite orbiting station to accept commands
  from Earth and relay data back to Earth.
• Base station(s)
• Ground stations responsible for accepting and
  relaying data to orbiting station.
• Carriers
• Long-range vehicles which carry smaller agents.
• Mappers
• Rovers responsible for mapping the geographical
  topography
• Local Positioning System (LPS) transmitters
• Ad-hoc local transmitters (arranged in a
  triangular configuration)used to construct
  accurate positional information for other agents.
  
• Mobile lab rovers
• Rovers which conduct experiments.
• Power sources
• Agents which generate/provide power (solar,
  geothermal, battery, etc.)
• Note that one physical device may be
  multi-functional, I.e., one physical device may
  be capable of performing more than one of the
  functions above.

9
Communication

• Ad-hoc networking. Ability to relay information
  in an ad-hoc network, with dynamic upload
  stations and minimal centralized control.
• Location awareness. Agents need to be aware of
  their positions relative to other agents, either
  through adjacent signal detection, or through a
  local positioning system (LPS).

10
Coordination

• Coordination of diverse elements. System must be
  able to coordinate agents with diverse functions
  (as listed in Types of Sensor Agents).
• Dynamic evolution of behavior. Multi-functional
  devices may dynamically change behavior, I.e.
  transform from a mapping agent to a lab agent
  dynamically.
• Staggered lifetimes. Different groups of agents
  may be activated at different times. One group's
  completion may trigger another group's
  activitation. Lifetimes of different groups may
  also be staggered to prolong mission life.
• Predictive coordination. System must adaptively
  use algorithms to coordinate multiple agents to
  meet conditions on-hand. For example, multiple
  agents may self-coordinate to explore a cave by
  forming an arm of the ad-hoc network into the
  cave.
• Environmental learning. System must collectively
  learn to avoid hazards, navigate the topography,
  zone in on interesting areas, etc.

11
Control

• Integrated remote/autonomous control. Since
  sensor networks may be deployed in locations not
  permitting real-time control, such systems must
  support autonomous control guided by high-level
  remote control commands.
• Safe bootstrapping. System must have capability
  to be reset to a safe operating state.
• Dynamic reconfiguration. System must have
  capability to be dynamically reconfigured by
  downloading new software.

12
Fault Tolerance/Survivability

• Minimal centralized control. Avoid centralized
  control of groups of agents to avoid a single
  point of failure.
• Power management. Methods to conserve power, and
  to efficiently distribute power, may prove
  essential.
• Fault recovery. The ability to recover from
  unseen obstacles, faults, accidents, etc, using
  techniques such as physical recovery, control
  reset, etc. Backup procedures when primary
  procedures fail are also essential.

13
Security

• Communications encryption. May want to encrypt
  communication if sensor network is deployed in
  enemy territory.
• Timed expiration. Software or hardware needs to
  self-destruct to avoid being stolen by enemies.

14
Example Scenario

• Orbiting station
• Base station(s)
• Carriers
• Mappers
• Local Positioning System (LPS) transmitters
• Mobile lab rovers
• Power sources

15
1. Orbiters deploy base stations
16
2. Base stations deploy carriers
17
3. Carriers deploy mappers and LPS transmitters
18
4. Mappers establish topography and areas of
interest

• Mappers coordinate through pheromone mechanisms
  and LPS coordinates
• Focus in on interesting areas
• Avoid hazards

19
5. Mappers return to carrier
20
6. Mobile lab rovers are deployed towards
interesting areas
21
7. Mobile lab rovers conduct experiments
22
8. Mobile lab rovers return to carrier
23
9. Carrier returns to base station, coordinates
with other carriers
24
10. Carrier pursues further sub-missions.
25
Important Properties

• Strength in Numbers Deploy a large number of
  miniature devices
• Fault tolerant
• More cost effective if devices are manufactured
  in large quantities
• Evolution
• Entire system evolves to adapt to conditions
• Devices change functionality
• Emergent Behavior
• Desired goal is achieved through cooperation of
  large number of autonomous devices which
  communicate among themselves