A Reputation and Trust Based Multi- Modal Sensor Network for Environmental Monitoring

Transcription

1 A Reputation and Trust Based Multi- Modal Sensor Network for Environmental Monitoring Edel O Connor Prof. Alan F. Smeaton, Prof. Noel E. O Connor & Prof. Dermot Diamond This Beaufort Marine Research Award is carried out under the Sea Change Strategy and the Strategy for Science Technology and Innovation ( ), with the support of the Marine Institute, funded under the Marine Research Sub-Programme of the National Development Plan Beaufort Research Awards in Marine Sensing

2 Presentation Outline Water management and issues with in-situ WSNs. Multi-modal sensor networks Data Aggregation Pilot studies River Lee water depth study Water level prediction for adaptive sampling Trust and reputation framework Beaufort Research Awards in Marine Sensing

3 Water Management Water management is an important part of the monitoring of the natural environment. For many years water managers relied on field measurements for coastal monitoring and water quality evaluation. However this process is being revolutionised through the introduction of new technologies such as sensor networks Image: John Cleary Beaufort Research Awards in Marine Sensing

4 Issues Current state of the art in chemo/biosensor networks not at a stage for reliable long-term large scale deployment. Even without the complexity of chemo-bio sensing, still considerable issues Sensors subject to harsh conditions Bio-fouling Limited spatial resolution Difficult to monitor large areas over long periods of time Unsuitable for certain environments and the immediate detection of certain events Developments in sensor research pushing towards ever cheaper systems Huge information overload user requires reliable event detection. Image: /images/ferryboxuse

5 Multi-modal sensor networks The incorporation of alternative sensing modalities such as visual sensors, alongside an in-situ WSN can help to overcome some of these problems. Beaufort Research Awards in Marine Sensing

6 Test Sites Requirements River Lee Galway Bay River Tolka Network x x Power x x Security x x Multiple sensing modalities x x Interesting from marine perspective x x x

7 Data Aggregation Camera data Beaufort Research Awards in Marine Sensing

8 Data Aggregation Satellite data Beaufort Research Awards in Marine Sensing

9 In-situ sensor data and context data Deploy: River Lee Rainfall Radar processing Image: (IDS: Intelligent Data Systems) SmartBay: Galway Bay Image: Marine Institute Beaufort Research Awards in Marine Sensing

10 Feet 00:10:43 02:22:03 04:33:23 06:44:42 08:56:02 11:07:22 13:33:17 River Lee Water Depth Study 15:44:36 17:55:56 20:07:15 22:18:35 00:29:54 02:41:14 04:52:34 07:03:53 09:15:13 11:26:33 13:37:52 15:49:12 18:00:31 20:11:51 22:23:10 00:34:30 02:45:50 04:57:10 07:08:29 09:19:49 11:31:09 13:42:29 15:53:48 18:05:07 20:16:27 22:27:46 Feet 00:02:07 00:45:53 01:29:40 02:13:27 02:57:13 03:41:00 04:24:46 05:08:33 05:52:19 06:36:06 07:19:52 08:03:39 08:47:26 09:31:12 10:14:59 10:58:45 11:42:32 12:26:18 13:10:05 13:53:51 14:52:14 15:36:00 16:19:47 17:03:33 17:47:20 18:31:06 19:14:53 19:58:39 20:42:26 21:26:12 22:09:59 22:53:45 23:37:32 Feet 00:14:32 00:58:18 01:42:04 02:25:51 03:09:37 03:53:24 04:37:10 05:20:57 06:04:43 06:48:30 07:32:17 08:16:03 08:59:50 09:43:36 10:27:23 11:11:09 11:54:56 12:38:42 13:22:29 14:06:15 14:50:02 15:33:48 16:17:35 17:01:21 17:45:08 18:28:54 19:12:41 19:56:27 20:40:14 21:24:00 22:22:22 23:06:09 23:49:55 Water Depth July 2008 Water Depth 03 July 2008 Water Depth 22 July Series1 Series1 15 Series Time Time Time Class Distance Error Classific ation Rate C 1 C 2 C

11 Multi-modal sensor networks adaptive sampling Beaufort Research Awards in Marine Sensing

12 Reputation and Trust-based multi-modal sensor network Development of a reputation and trust-based multi-modal sensor network Adaptation of a model developed for in situ sensor networks known as RFSN (RFSN Ganeriwal & Srivistava 2008). Adaption of this model to multi-modal sensor networks Beaufort Research Awards in Marine Sensing

13 Community of Sensor s j k i l m

14 Community of Sensor s j k i l m

15 RFS N RFS N RFS N RFSN j RFSN RFSN RFSN k l i RFSN m

16 Community of Sensor s RFSN [j,k,l,m] WATCHDOG Series of outlier detection protocols, outputs cooperation metrics for each of the nodes. Cooperation metrics REPUTATION Updates reputation for each of the nodes [I, j, k,l] i.e. R ij, R ik, R il, R im trust External evidence

17 Community of Sensor s i j k m l

18 Community of Sensor s RFSN j

19 Community of Sensor s RFSN [j,k,l,m] WATCHDOG Series of outlier detection protocols, outputs cooperation metrics for each of the nodes. Cooperation metrics REPUTATION Updates reputation for each of the nodes [I, j, k,l] i.e. R ij, R ik, R il, R im trust External evidence

20 Acknowledgements Beaufort Research Awards in Marine Sensing Marine Institute CLARITY Centre for Sensor Web Technologies, Science Foundation Ireland under grant 07/CE/I1147 HRDDS (GHRSST Project) and David Poulter at the National Oceanography Centre, Southampton, UK Beaufort Research Awards in Marine Sensing