AIT Chapter 5 Sensor Networks

Focus on interacting with the environment and not with humans

1. id-centric for MANET and 2. data-centric for WSN improving efficiency

1. Process information via sensing and actuation

2. Send answers instead of only numbers

3. Depends on application specifics

4. In-network event detection and processing

5. Combine WSNs with LIDAR, RADAR and web technologies

6. Must be fault-tolerant with a lifetime

7. Flexibility in form of programmability and maintainability

1. In case of wildfire to derive temperature map

2. Observe wildlife

3. Memory and energy constraint

4. Observe buildings for energy waste reduction or mechanical stress

5. Observe permafrost, glacier, landslide and trees

6. Facility management with intrusion and leakages

7. Machine surveillance and preventive maintenance

8. Precision agriculture

9. Medicine and health care

10. Logistics with equipping goods with sensor node

11. Telematics

12. Smart homes and cities

1. Behavior of wild rats in their natural habitat

2. Limited radio range therefor sporadic connectivity

1. Sensor-based landslide early warning system (SLEWS)

2. Volcano Deployment

3. Swiss experiment for micro-climate

1. Event detection and classification

2. Periodic measurement

3. Function approximation of space and or time

4. Edge detection in a function

5. Tracking of object

1. Core as WSN node

2. More additional equipment

IEEE 802.11n with higher data rates

1. (Micro-)controller

2. Memory

3. Sensors and actuators

4. Communication devices as transceiver

5. Power supply

Energy and size limits capabilities

1. Coordinating other components

2. Optimized for embedded applications

3. Low power consumption

Texas Instruments MSP430

1. 16-bit RISC core

2. 8-25 MHz

3. 2-16 kByte RAM

4. RT clock

5. Inexpensive

1. Interface on bit, byte or packet level

2. Frequency range on 433 MHz – 2.4 GHz on ISM band // Industrial, scientific, medical – license free bands

3. Power consumption to send/receive data?

1. RFM TR1000 family and 2. Chipcon CC 2400

1. Passive or active

2. Directed or omidirectional

3. Area of coverage

1. Primary not rechargeable batteries

2. Secondary rechargeable batteries with energy harvesting

1. Zinc-air with 3780 J/cm³

2. Lithium with 2880 J/cm³

3. Alkaline with 1200 J/cm³

1. Lithium with 1080 J/cm³

2. NiMHd with 860 J/cm³

3. NiCd with 650 J/cm³

1. Easy with laptop

2. Difficult to scavenge energy from environment

1. Light via solar cells with 10myeW/cm² to 15mW/cm²

2. Temperature gradients with 80myeW/cm² at 1V at d5K

3. Vibrations with 0.1 to 10000myeW/cm³

4. Pressure variation 330ymeW/cm³

5. Air or liquid flow

6. Acoustic noise

7. Passive human powered systems

8. Nuclear reaction with 80mW/cm³

1nJ with 1J=1Ws with 10⁹ instructions ~ 11500 per second for one day

Instruction cheaper than communication

1. Active

2. Idle

3. Sleep with variations down to 0.3 myeW // No timer at deepest level

1. Transmit

2. Receive // Major energy consumption

3. Idle

4. Sleep

1. Contention-based MAC protocols: always on

2. TDMA-based MAC protocols: synchronization OH

3. Help of dedicated receiver

P_amp = a_amp + b_amp*P_tx with P_tx radiated power and constants

peta=P_tx/P_amp

t=n/R*R_code with nominal data R and coding rate R_code

E_tx=T_start*P_start+ t*(P_txElec + P_amp) with average power P_start

1. Simple idea says switch greedily to lower mode

2. Better if paying attention to E_saved > E_{overheadAndTime}

3. Dynamic voltage scaling (DVS)

Power consumption P depends on clock frequency f and square of supply voltage v²

1. Randomly distributed

2. Regular planned and fix

3. Mobile active or passive