May 2004 doc 15 04 0228 01 004

May 2004 doc. : 15 -04 -0228 -01 -004 A Project: IEEE P 802. 15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: IEEE 802. 15. 4 Relative Location Date Submitted: May 2004 Source: Neiyer Correal, Frederick Martin, Motorola, Inc. Contact: F. Martin, Motorola, Inc. , 8000 W. Sunrise Blvd. Plantation, FL 33322 Voice: +1 954 -723 -6395, FAX: +1 954 -723 -3712, E-Mail: f. martin@motorola. com Re: Technical Contribution to TG 4 A Abstract: Relative location based on received signal strength ranging can be used for 2 dimensional location in personal area networks. Purpose: To provide information on potential location performance of networks based on IEEE 802. 15. 4. Notice: This document has been prepared to assist the IEEE P 802. 15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P 802. 15. Submission 1 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Motivation for Relative Location Techniques • Device data must be accompanied by its location – Actuator must know where to act – Alarm must be localized to be useful to a human operator • Placing & locating every device is labor-intensive • Cost of device location must be considered – GPS cost, power consumption can be prohibitive • Only an outdoor solution – Local Positioning Systems (LPS) need intensive installation • High Cost of wiring • Labor intensive Submission 2 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location • Goal: Self-location with few known-location nodes – Without wired infrastructure – Applicable to any environment • Difficulties – Ad hoc connectivity Advantages – Reference devices are the ‘infrastructure’ – Peer-to-peer range measurement possible – Accuracy/range extension • C 9 central computer B A 6 7 5 data link 8 3 4 2 1 Wireless Sensors ‘Reference’ ‘Blindfolded’ Submission 3 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location • • Conventional Location – Devices are located only with respect to fixed base stations z. A Relative Location – All devices calculate the distance to their neighbors central computer d 9, A d d d 9, B data link z. C d 9, C v z 12 d d Synch / data d Legend d d z 9 d d data link d Neu. RFons ‘Reference’ Architectural Blueprint zn: devices Architectural Blueprint da, b: range estimates ‘Blind’ Characteristics Long path lengths Complex installation Submission d d d z. B d central computer Benefits Higher rates Location Accuracy/Range Extension 4 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location in 1 Dimension Submission 5 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location – Physical Analogy • Relative location analogy: – Nails represent absolute location information – Springs’ natural lengths are the estimated range between nodes Relative Range Data Reference and Blind Neu. RFons Submission Location Mapping Algorithm – Spools are nodes Spools will move in the direction of force until sum of forces equals zero 6 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Range Estimation Techniques • Various media are possible RF Infrared Ultrasound Acoustic Signal Strength Techniques Time-Based Techniques • • Time Difference of Arrival (TDOA) relies on time synch – (ns) time-synch unlikely Time-of-Arrival (TOA) can be measured via round-trip At 1 meter range resolution for 3 ns time resolution (RF signal), wide bandwidth signal is needed. Submission • • 7 Received Signal Strength (RSS) is considered coarse due to the multipath channel Range estimate variance increases with range WE WILL FOCUS ON RSS ON RF SIGNALS Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location Simulation Path Loss Errors of s = 6 d. B Single room, 20 m by 10 m 3, 4, or 6 Reference devices 1 to 10 Blindfolded devices Reference devices located in corners Other devices placed at random • • Submission 8 Simulation of Indoor Relative Location – Signal Strength as a Ranging Technology Location Accuracy improves with either – More Reference devices – More Blindfolded devices. Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location Experiment Set-up • Measurements – Mot. Labs Plantation FL, office environment – 13 by 15 m area – 44 devices (0. 2 /m 2) – 4 reference devices in corners of area – Nodes at 1 m height – Multipoint to multipoint – 44*43*5 = 9460 measurements – 2. 4 GHz signal, 40 MHz BW Figure: Area Map, Device # and Location Submission 9 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location Experiment Set-up Key: #T True Location #R Relative Location Estimate • • RMS Location Error of 2. 14 m Histogram of Errors RMS is 15% of Ref. device separation Submission 10 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location Radio Testbed • Testbed Device – 900 MHz, 50 kbit/s FM – 8 channels – Sensor Suite • RSS Ranging Submission 11 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Relative Location Measurement Results • • Environments – Outdoor Parking Lot – Residential Home 9 m x 9 m area, 4 by 4 grid – Reference nodes at corners • RMS location errors – 1. 0 m in parking lot – 2. 1 m in house Above: Map of the Perkins home and grid of 16 device locations. Side: Parking lot experiment. Devices are located on top of blue upside-down recycling bins. Submission 12 Correal, Motorola, Inc.

May 2004 doc. : 15 -04 -0228 -01 -004 A Summary 802. 15. 4 can be used for location under some conditions RSSI appears to be a viable method for location resolution Cooperative location can enhance location estimation -- higher density of nodes results in improved location estimation For reference node separation on order of 10 m, 2 D location with RMS errors on order of 2 m is possible For more information, see N. Patwari, A. O. Hero, III, M. Perkins, N. Correal, R. J. O’Dea, “Relative location estimation in wireless sensor networks, ” IEEE Trans. On Signal Processing, vol. 51, no. 8, August, 2003, pp. 2137 -2148. Submission 13 Correal, Motorola, Inc.