1. 1 Simultaneous Localization and Mobile Robot Navigation in a Hybrid Sensor Network Suresh Shenoy and Jindong Tan Michigan Technological University Intelligent Robots and Systems (IROS 2005).

2. 2 Outline Introduction Simultaneous Localization & Navigation Simulation Conclusions

3. 3 Introduction The static sensors around the event may request the mobile sensor/robots to navigate themselvesThe static sensors around the event may request the mobile sensor/robots to navigate themselves The location information is important for the navigation of mobile robotsThe location information is important for the navigation of mobile robots

4. 4 Introduction Localization algorithmLocalization algorithm Range basedRange based Add additional hardware (e.g: GPS)Add additional hardware (e.g: GPS) Range-free basedRange-free based Location information can be obtainedLocation information can be obtained RSSIRSSI Time of arrival or time difference of arrivalTime of arrival or time difference of arrival

5. 5 Assumption The mobile robots are equipped with GPSThe mobile robots are equipped with GPS The mobile robots are equipped with more resource such as sensor, power, computationThe mobile robots are equipped with more resource such as sensor, power, computation

6. 6 Goal N3N3 N8N8 N6N6 N 12 N 15 N2N2

7. 7 Mobility Based Range-Free Localization A (x,y) Bounding box B.box = {(x, y, range) : (xmin, xmax, ymin, ymax)} Communication range

8. 8 Mobility Based Range-Free Localization A (x,y) B.box = {(xmin, xmax, ymin, ymax)∩(xmin ’, xmax ’, ymin ’, ymax ’ )} (x ’,y’)

9. 9 Simultaneous Localization & Navigation

10. 10 Communication of Navigation N3N3 N8N8 N6N6 N 12 N 15 N2N2 Event Packets 115 12 6 8 2 3 6 Event Node’s traversed No of nodes

11. 11 Communication of Navigation N3N3 N8N8 N6N6 N 12 N 15 N2N2 Navigation Request packet: broadcast to its one hop neighbor Location Packet: location of the mobile robot Combined

12. 12 Communication of Navigation N3N3 N8N8 N6N6 N 12 N 15 N2N2 Navigation Packets: the acknowledgement packet that contains the bounding box

13. 13 Mobile Robot Navigation in Sensor Networks – situation 1 N9N9 N3N3 N2N2 N8N8 Navigation Packet A random walk strategy is adopted

14. 14 Mobile Robot Navigation in Sensor Networks – situation 1 N9N9 N3N3 N2N2 N8N8 N1N1 N7N7 N 1, N 9, N 3, N 7 N 2 is neighbor of N 9 and N 3

15. 15 Mobile Robot Navigation in Sensor Networks – situation 1 N9N9 N3N3 N2N2 N8N8 N1N1 N7N7 The robot always moves along the longer side of the bounding box

16. 16 Mobile Robot Navigation in Sensor Networks – situation 1 N9N9 N3N3 N2N2 N8N8 N1N1 N7N7

17. 17 Mobile Robot Navigation in Sensor Networks – situation 1 N9N9 N3N3 N2N2 N8N8 N1N1 N7N7 If the robot realizes that it has lost communication with the sensor node, it hence returns to a previous location

18. 18 Mobile Robot Navigation in Sensor Networks – situation 2 N3N3 N9N9 N 10 N8N8 N2N2

19. 19 The robot navigation algorithm

20. 20 Simulation Using ns-2 and matlabUsing ns-2 and matlab Network topology: 1000 * 1000Network topology: 1000 * 1000 Random distribution of nodesRandom distribution of nodes Communication range: 40mCommunication range: 40m The number of sensor nodes: 100The number of sensor nodes: 100

21. 21 Simulation

22. 22 Simulation – navigation path taken by the mobile robot

23. 23 Simulation

24. 24 Conclusions Navigating the mobile robot using sensor nodes that are not localized in an event driven network, which in turn get localized as the robot is navigatedNavigating the mobile robot using sensor nodes that are not localized in an event driven network, which in turn get localized as the robot is navigated

25. 25 Thank you!!