Wireless Sensor Networks Powered by Ambient Energy Harvesting (keynote).ppt
- 1. WSN-HEAP Wireless Sensor Networks Powered by Ambient Energy Harvesting Winston Seah ©2009
- 2. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 3. What are WSNs? Wireless Sensor Networks Originated from military/security applications, many new potential applications have emerged in areas such as medical, industrial, automotive, agriculture, environmental and structural health monitoring Consists of sensor nodes distributed over an area monitoring some phenomena Sensors monitor temperature, pressure, sound, vibration and motion Typically powered by on-board batteries MICAz mote IRIS mote
- 4. Old Assumptions Deployed randomly, e.g. air dropped Operational lifetime limited by battery Densely deployed to provide redundancy No concern for environmental implications caused by hardware, especially batteries Predominantly driven by military and/or short-term surveillance oriented applications Communications subsystem design is driven primarily by need to extend the limited battery lifetime
- 5. New Applications Structural Health Monitoring – monitoring bridges, tunnels, dams, ancient monuments, construction sites, buildings, roads, railways, land masses, etc. Agriculture and food industry – environmental monitoring, precision agriculture, facility automation (greenhouse control, animal-feeding system), etc Industrial automation – M2M-based machine and process control Building automation, smart homes, smart offices, smart spaces Environmental monitoring for conservation
- 6. Structural Health Monitoring Compelling need for SHM because Earthquakes can shake buildings, even in Singapore (e.g. Sumatran earthquakes) Soil movement from construction and excavation works may cause buildings to become unstable (e.g. MRT/subway Tunneling Works) April 2004
- 7. Structural Health Monitoring Compelling need for SHM because Structures may weaken over time (e.g. bridges, building foundations, elevated roads) due to bacterial, chemical, or (sea) water damage Wear-and-tear may result in structural deformation and mechanical faults (e.g. bridges, railway tracks, etc.)
- 8. Deficiencies of current SHM approaches Sensors welded / embedded into critical structures Infeasible / hazardous to replace / recharge batteries Sensors are wired to data loggers (sinks) Cabling is expensive, messy, prone to damage, hazardous, non-recyclable and has limited coverage Offline data collection (non real-time) Early warning signals may not be detected in time
- 9. WSN for SHM Why use WSN? Prevalent transmission technology IEEE 802.15.4, 802.11, 802.15.1 Higher availability and wider coverage Reduced costs and wastage Typical wiring costs US$130-650 per metre Wireless tech can eliminate 20-80% of costs Reduce interferences from electrical sources Less vulnerable to disruptions arising from cable damage
- 10. WSN for Agriculture Lofar Project (NL) - WSN for Potato farming (TH) Grape Networks (US) SoilWeather (FI) CSIRO (AU)
- 11. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 12. Energy Harvesting Power has been and remains the key WSN issue Alternative source of energy for WSNs Gather energy that is present in the environment, i.e. ambient energy Convert the energy into a form that can be used to power devices Assumes energy source is well characterized, regular and predictable Energy scavenging refers to scenarios where energy source is unknown and highly irregular
- 13. Energy Harvesting for SHM Why Ambient Energy Harvesting? Batteries in sensor nodes embedded in structures are not easily replaceable No danger of battery leakage (corrosive to structure) and environmentally-friendly Operate perpetually without need for human intervention Can be used in emergencies when power supply is not available
- 14. Energy Harvesting for Agriculture Why Ambient Energy Harvesting? Batteries in sensor nodes in plantation are not easily replaceable high risk of damaging crops No batteries no danger of battery leakage and polluting the environment Operate perpetually without need for human intervention
- 15. Energy Harvesting for WSN usage Mechanical (Vibration or Strain) energy harvesters Bridges, roads, railway tracks movement Trains and vehicles cause vibration Solar films Thin solar films that can be “pasted” on buildings are becoming a reality Ambient light can also be harvested Water Mini/Micro-hydroelectric generators in irrigation canals, streams, rivers, waterways, pipes, etc.
- 16. Energy Harvesting for WSN usage Ambient airflow Besides natural airflow, wind is also generated by movement of vehicles, and even air conditioning Ambient RF Available everywhere (e.g. from cell phones, WiFi) 8 µW to 420 µW (IEEE Trans on Power Electronics, May 2008) Pressure Energy is generated due to pressure (e.g. from movement of people)
- 17. Batteries vs Supercapacitors Batteries Limited Recharge cycles Higher storage density (30-120 Wh/kg) Environmentally unfriendly and prone to leakage Capacitors/Supercapacitors Virtually unlimited recharge cycles Capacitors have lower storage density than batteries (0.5-10 Wh/kg) Supercapacitors have potentially higher energy storage density than batteries/capacitors (30-300 Wh/kg)
- 18. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 19. WSN-HEAP Acronym for Wireless Sensor Networks Powered by Ambient Energy Harvesting Used for denoting WSNs that are solely powered by energy harvesting devices using capacitors/supercapacitors excludes WSNs that use energy harvesters to supplement battery power
- 20. WSN-HEAP node
- 21. Energy Model of WSN-HEAP node Energy harvesting is only energy source Different energy harvesting (charging) rate across time and physical domains Average energy charging rate is lower than the rate of energy consumption Short duty cycle
- 22. Major Research Groups UCLA CENS: Heliomote Energy Harvesting System EPFL Sensor Scope Project UC Berkeley WEBS (Wireless Embedded Systems) Heliomote by UCLA EPFL UC Berkeley
- 23. Sensor Nodes with Energy Harvesting Research Heliomote (V. Raghunathan et. al., IPSN 2005) AmbiMax (C. Park and P. H. Chou, SECON 2006) Trio (P. Dutta et. al, IPSN 2006) Heliomote AmbiMax Trio Mote
- 24. Sensor Nodes with Energy Harvesting Research Piezoelectric Igniter (Y. K. Tan and S. K. Panda, IEEE ICIT 2006) Everlast (F. I. Simjee and P. H. Chou, IEEE Trans. on Power Electronics, 2008) Everlast Piezoelectric Igniter
- 25. Sensor Nodes with Energy Harvesting Commercial Ambiosystems Microstrain Enocean Crossbow Solar-powered sensor node by Enocean Energy converter for linear motion by Enocean Battery-less motes by Ambiosystems Solar-powered sensor node by Microstrain Solar-powered (supplemented) sensor node by Crossbow
- 26. Current State-of-the-Art Energy Harvesting Rates Technology Power Density (µW/cm2 ) Energy Harvesting Rate (mW) Duty Cycle (%) Vibration – electromagnetic 4.0 0.04 0.05 Vibration – piezoelectric 500 5 6 Vibration – electrostatic 3.8 0.038 0.05 Thermoelectric 60 0.6 0.72 Solar – direct sunlight 3700 37 45 Solar – indoor 3.2 0.032 0.04 Power consumption for MICAz sensor node is 83.1mW in the receive state and 76.2mW in the transmit state. Source: B. H. Calhoun et. al., “Design Considerations for Ultra-Low Energy Wireless Microsensors Nodes”, IEEE Transactions on Computers, Vol. 54, No. 6, June 2005
- 27. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 28. Research Challenges WSN Architecture Power Management Modulation and Coding Medium Access Control (MAC) Schemes Routing Protocols Transport Protocols
- 29. WSN Architecture Single-Hop Single-Sink Architecture used by most WSNs with energy harvesters
- 30. WSN Architecture Multi-Hop Single-Sink Architecture used by many WSNs with on-board batteries
- 31. WSN Architecture Multi-Hop Multi-Sink Increases network capacity
- 32. Power Management Most work on power management in WSNs using energy harvesting devices is done by M. Srivastava’s group in UCLA ISLPED 2003, SIGMETRICS 2004, IPSN 2005, DAC 2006, ISLPED 2006, ACM TECS 2007 Their main focus is on estimating amount of energy that can be harvested in future to optimize duty cycles and scheduling of tasks Main assumption is that harvested energy is used in conjunction with battery power Their energy model is different from ours
- 33. Challenges in Power Management in WSN-HEAP In WSN-HEAP, higher transmission power means longer energy harvesting time Reduced sending rate However, higher transmission power also means that there are more potential awake neighbors to forward data packets to What is the optimal transmit power to maximize throughput?
- 34. Modulation and Coding IEEE 802.15.4 Most commonly assumed physical data transmission standard for sensor networks Commonly referred to as Zigbee Used in many popular sensor motes (e.g. MICAz, TelosB) IEEE 802.11 Widely used for WLANs Not power-efficient Used in some WSN applications Desired features of schemes for WSN-HEAP Need to be more opportunistic Quick to transmit and fully utilize the limited energy
- 35. Sensor MAC protocols S-MAC (W. Ye, Infocom 2002) Periodic sleep and wakeup cycles Latency is increased as a result Variants include T-MAC and DSMAC to improve performance under specific scenarios B-MAC (J. Polastre, SenSys 2004) Adaptive preamble sampling scheme to reduce duty cycle and minimize idle listening
- 36. Sensor MAC protocols TRAMA (V. Rajendran, SenSys 2003) TDMA-based algorithm Time synchronization is required Sift (K. Jamieson, EWSN 2006) Designed for event-driven WSN to minimize collisions when event occurs
- 37. Challenges in MAC for WSN-HEAP Difficult to use TDMA Time synchronization is harder in WSN-HEAP than conventional WSNs Difficult to use sleep-and-wakeup schedules Not possible to know exactly when each node is awake Difficult to set duty cycles Energy harvesting rates change with time and place
- 38. Routing Protocols Flat routing Directed Diffusion (C. Intanagonwiwat, Mobicom 2000); Solar-aware Directed Diffusion (T. Voigt, LCN 2003) Variants include Rumor Routing, Gradient-Based Routing (GBR), Random Walks Hierarchical Routing Makes use of clustering and data aggregation LEACH (W. Heinzelman, HICSS 2000) Variants include PEGASIS, TEEN, APTEEN
- 39. Routing Protocols Geographic Routing GeRaF (M. Zorzi, IEEE Trans on Mobile Computing, 2003) GPSR (B. Karp, MOBICOM 2000) Variants include GAF, GEAR, SPAN
- 40. Challenges in Routing for WSN-HEAP Difficult to determine next-hop neighbor Not possible to determine exact wakeup schedules Many sensor routing protocols assume knowledge of neighbors Complete routes may not be available when a data packet is sent Delay-Tolerant Networking (DTN) may be a solution but be adapted to WSN-HEAP
- 41. Challenges in Routing for WSN- HEAP How to efficiently route data in WSN-HEAP when different nodes have different energy harvesting rates? How to aggregate or disseminate sensor data in WSN-HEAP?
- 42. Transport Protocols Variable Reliability STCP (Y. G. Iyer, ICCCN 2005) Event-based Reliability ESRT (Y. Sankarasubramaniam, MobiHoc 2003) Congestion Control Flush (S. Ki, Sensys 2007) CODA (C.-Y. Wan, Sensys 2003) Fusion, CCF, PCCP, ARC, Siphon, Trickle
- 43. Challenges in transport protocols for WSN-HEAP How to detect congestion when a node is only awake for short periods of time? How to send the feedback from the sink to the source node when we do not know exactly when the source node would be awake? How to provide fairness if there are nodes with different energy harvesting rates?
- 44. Technical Challenges Not possible to know exactly which is the awake next- hop neighbor to forward data to Not possible to predict exactly when the node will finish harvesting enough energy
- 45. WSN-HEAP vs battery-operated WSNs Battery-operated WSNs Battery-operated WSNs with energy harvesters WSN-HEAP Goal Latency and throughput is usually traded off for longer network lifetime Longer lifetime is achieved since battery power is supplemented by harvested energy Maximize throughput and minimize delay since energy is renewable and the concept of lifetime does not apply Protocol Design Sleep-and-wakeup schedules can be determined precisely Sleep-and-wakeup schedules can be determined if predictions about future energy availability are correct Sleep-and-wakeup schedules cannot be accurately predicted Energy Model Energy model is well understood Energy model can be predicted to high accuracy Energy harvesting rate varies across time, space as well as the type of energy harvesters used
- 46. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 47. Application Examples Self-powered railway sleeper monitoring system Stability Monitoring of Bridges and Expressways
- 48. Wireless Monitoring Systems for Rail Systems Railway track and bridge monitoring Remote (wireless) rail temperature preventive maintenance system in UK’s high speed rail network since 2005 Next-generation wireless mesh for predictive maintenance demonstrated for Network Rail (UK) in 2007 Battery-powered Requires human intervention for battery replacement Poses safety issues and may disrupt rail operations
- 49. Self-Powering (Ambient Energy Harvesting) Vibrational energy from track deflections Wind energy from passing trains in tunnels Solar energy for outdoor tracks
- 50. Self-Powered, Online Rail-track Sleeper Monitoring Benefits of wireless Mature and prevalent technology WiFi, ZigBee Higher availability and wider coverage Reduced costs and wastage Online monitoring and remote control Self-Powered, Wireless Monitoring Instrument (vibration, solar) on sleepers on viaduct and at-grade stations Benefits of self-powering – Sustainable – Environmental friendliness – Economical – Safety – Commercially available
- 51. Stability Monitoring of Bridges and Expressways using WSN- HEAP Pasir Panjang Semi-Expressway Photo Source: SysEng (S) Pte Ltd
- 52. Photo Source: SysEng (S) Pte Ltd
- 53. Ongoing Research MAC Protocols for WSN-HEAP Adapt and compare different MAC protocols for use in WSN-HEAP Design MAC scheme for WSN-HEAP Validated analytical and simulation results; working on experimentation Results enable network designers to determine the suitable MAC protocol to use to maximize throughput given the average energy harvesting rates and the number of WSN-HEAP nodes to deploy
- 54. Ongoing Research Routing and Node Placement Algorithms Different node placement schemes affect network performance Optimal choice of a node placement scheme and routing algorithm is crucial in maximizing goodput Empirical Characterization Energy harvesting rates Link quality measurements Packet delivery ratios
- 55. Lab Feasibility Study (Solar)
- 57. Empirical Characterization MSP430 microcontroller & CC2500 radio transceiver by TI Indoor solar energy harvester is provided by Cymbet Thermal Energy Harvester by Micropelt
- 59. Outline Quick Introduction of Wireless Sensor Networks (WSN) Energy Harvesting for WSN WSN-HEAP Research Challenges Application Examples and Ongoing Research Concluding Remarks
- 60. Conclusions and Future Work WSN-HEAP are viable solutions to making WSN more pervasive Increase the commercial viability of wireless sensor networks since maintenance costs are reduced. Since energy harvesters make use of energy that is otherwise wasted, WSN-HEAP contribute to environmental sustainability Increased structural monitoring capabilities will lead to more early warnings, thereby reducing the risk of deaths or injuries when structures collapse
- 61. Conclusions and Future Work Focus on maximizing throughput/goodput and minimizing delays given the amount of energy that we can harvest from the environment. Amount of sensor data should increase when energy harvesting rates increase and number of sensor nodes increase Reliability issues are important in some sensor network applications Set up a testbed to validate our ideas and protocols.
- 62. Thank you. For more information: Email: winston.seah@outlook.com
Editor's Notes
- #6: * * Bacterial Damage - Bacteria can produce hydrogen sulfide, which oxidizes into the corrosive sulfuric acid, harming the surface of concrete. However, calcium carbonate in limestone can make the concrete more acid resistant. * Chemical Damage - The over use of substances that contain too much acid, such as sulfates, chlorates, and certain carbonates increase the pH of the concrete and can lead to cracking and dissolving. This is a serious issue with construction, as too much damage can cause a building to collapse. * Sea Water Damage - A particularly harmful effect on concrete structures is the erosion that occurs when it comes into contact with running water, such as bridges, where over the years, a bridge can collapse because of the amount of undersea damage sustained. Damage can be done to cars, people could get hurt, and traffic would be insane in an event such as this.
- #7: * * Bacterial Damage - Bacteria can produce hydrogen sulfide, which oxidizes into the corrosive sulfuric acid, harming the surface of concrete. However, calcium carbonate in limestone can make the concrete more acid resistant. * Chemical Damage - The over use of substances that contain too much acid, such as sulfates, chlorates, and certain carbonates increase the pH of the concrete and can lead to cracking and dissolving. This is a serious issue with construction, as too much damage can cause a building to collapse. * Sea Water Damage - A particularly harmful effect on concrete structures is the erosion that occurs when it comes into contact with running water, such as bridges, where over the years, a bridge can collapse because of the amount of undersea damage sustained. Damage can be done to cars, people could get hurt, and traffic would be insane in an event such as this.
- #9: wires, once used, cannot be reused.
- #13: wires, once used, cannot be reused.
- #14: wires, once used, cannot be reused.