Impact of mobility and MAC scheduling on positioning with UWB BAN - PowerPoint PPT Presentation

Impact of mobility and MAC scheduling on positioning with UWB BAN Cooperative Communications in Body Area Networks CORMORAN project - ANR Arturo Guizar Claire Goursaud Jean-Marie Gorce

Context WSN vs WBAN -> different localization constraints • On-Body, Body-to- Body, … Figure 1: Motion Capture and Navigation Group Scenarios

Context WSN vs WBAN -> different localization constraints • On-Body, Body-to- Body, … Figure 1: Motion Capture and Navigation Group Scenarios

Related Work There are no proposals for MAC design considering: • MAC scheduling impact on positioning estimation • under realistic mobility for WBAN scenarios

Context 1. Introduction 2. System Model - Network Topology - Localization Assumptions 3. Ranging estimation constraints - 3-WR packets delays - MAC strategies 4. MAC Scheduling for accurate positioning - Mobility Model - Simulation parameters - Results: Impact on positioning estimation 5. Scheduling for Motion Capture - Simulation parameters - Results: Performance for Motion Capture 6. Conclusion

Network Topology Mesh IR-UWB WBAN - full connectivity [Hamie2013] on-body mobile nodes (i = 1 … N ) on-body anchor nodes (j = 1 … M) N T = N + M A set of anchors define a Local Coordinate System (LCS) to localize nodes under mobility. Figure 3: Yoga Scenario

Network Topology Mesh IR-UWB WBAN - full connectivity [Hamie2013] on-body mobile nodes (i = 1 … N ) on-body anchor nodes (j = 1 … M) N T = N + M A set of anchors define a Local Coordinate System (LCS) to localize nodes LCS under mobility. Figure 3: Yoga Scenario

Localization Assumptions Ranging Three-Way Ranging (3WR) protocol • Request (Q ij ) • Response packets (R ij and S ij ) Positioning Time Difference of Arrival (TDOA + LLS) technique for Positioning estimation P’ i (t) [Xiao2010] • j ≥ 4 anchors for distributed localization (multilateration) Figure 4: Ranging estimation with 3WR

Ranging estimation constraints 1) Impact on ranging: 1 mobile node → 1 anchor, using TOA In our previous work [Guizar2014], we show that the time to send the first response Δ t1 has more impact on ranging estimation than the time taken for the second response Δ t2. Arturo Guizar, Anis Ouni, Claire Goursaud . "Impact of Mobility on Ranging Estimation using UltraWideband, " ACM SIGCOMM 2014 conference, Aug 2014, Chicago, United States.

Ranging estimation constraints 1) Impact on ranging: 1 mobile node → 1 anchor, using TOA 2) Impact on Positioning: 1 mobile node → M anchors, with TDOA 3) Impact on Posture: N mobile nodes → M anchors, set of positions Figure 5: Error on Ranging estimation with TDOA (static node vs mobile node)

Scheduling for accurate positioning First contribution: Scheduling strategies to increase the positioning accuracy of 1 mobile node in a WBAN Network Topology: 1 mobile node → M anchors, using TDOA We introduce the Broadcast case on the Request Packet for the Best Strategy of the study All requests first Ordered transaction Three period order Priority for Response 1

Simulation tools and parameters Discrete-event simulation approach using the WSNet simulator http://wsnet.gforge.inria.fr/ • Physical (PHY) Layer based on the IEEE802.15.6 PHY UWB • Default mode (OOK modulation, data rate 0.4875 Mbps) [Niemelä2013] • Line of Sight (LOS) channel without packet loss. • The radio is capable detect the 1st path of IR-UWB • Medium Access Control (MAC) Layer based on the TDMA protocol and we assume that it is beacon enabled. Impact on positioning estimation with the Root Mean Square Error (RMSE): N f is the number of frames during the simulation P ref i (t) is a reference position

Mobility Model Realistic scenario - CORMORAN measurement campaign 2014 Human body movements have been captured on an area of 10 x 6 m using a system based on a Vicon Mx40 at 100Hz. Vicon -> PyLayers Mobility Model -> WSNet node's position http://pylayers.github.io/pylayers/

Impact on positioning estimation Parameters of simulation: • 1 mobile node (right wrist) and 4 anchors • Realistic (Right Front Chest, Left Front Chest, Left hip, Back Chest) Speed Factor x(1-10) to accelerate the Human body movement • We compare P’ i (t) with the real position at the end of the frame P’ final i (t)

Impact on Positioning estimation S1: All request first S3: Three period order S2: Ordered transaction S4: Priority for Response 1 - S1 and S2 give better positioning estimation than S3 and S4 - B-S1 reduce error on positioning and delay (Gain Broadcast = (M-1)*N , 3 slots in our case)

Discussion

Other ranging issues • Impact of 3-WR scheduling for Motion Capture • Impact of the nodes resource allocation • Quantifying the ranging error related to the mobility and the channel effects • 3-WR delays with aggregated & cooperative strategies

References A. Guizar, A. Ouni, C. Goursaud, C. Chaudet, and J. Gorce , “Quantifying the Impact of Scheduling and Mobility on IR-UWB Localization in Body Area Networks,” in 12th IEEE International Conference on Body Sensor Networks, MIT, Cambridge, USA, 2015. A. Guizar, A. Ouni, C. Goursaud, N. Amiot, and J. Gorce, “Impact of MAC scheduling on positioning accuracy for motion capture with UWB body area networks,” in 9th International Conference on Body Area Networks, London, 2014. B. Denis, N. Amiot, B. Uguen, A. Guizar, C. Goursaud, A. Ouni, and C. Chaudet, “ Qualitative Analysis of RSSI Behavior in Cooperative Wireless Body Area Networks for Mobility Detection and Navigation Applications,” in 21st IEEE International Conference on Electronics Circuits and Systems, (Marseille, France), Dec. 2014. A. Guizar, A. Ouni, C. Goursaud, and J.M. Gorce, “Proposition d’ordonnancement pour une meilleure localisation de nœuds mobiles dans les Réseaux Corporels sans fils avec ULB, “ In 17èmes Rencontres Francophones sur les Aspects Algorithmiques de Télécommunications (AlgoTel), Baune, France, 2015 A. Guizar, A. Ouni, and C. Goursaud, “Impact of mobility on ranging estimation using Ultra Wideband, “ in Proceedings of the Fourth Networking Networking Women Workshop

Ongoing Work

Thanks for your attention Arturo M. J. Guizar http://www.pylayers.org/cormoran

Localization constraints Among all the problems … • Physical Layer • Synchronization • Data Stream • Maximize Channel Utilization • Real-Time • Off-line analysis applications My main interest: • Protocol design at MAC layer • Network configuration • QoS (latency, precision) Figure 2: Ranging estimation with UWB