**10. Conclusion**

56 Will-be-set-by-IN-TECH

Normalized measured impulse response

Normalized magnitude after background subtraction

Target range estimates

0 20 40 60 80 100

Time of flight [ns] **Figure 58.** Target echo detection - measured impulse response (blue), normalized signal magnitude after background subtraction (green), CFAR test statistic (red), CFAR adaptive threshold (cyan), indices of detected targets by CFAR (magenta) and Gaussian mixtures representing the estimated target

as a range estimate that corresponds to the same target. When multiple range estimates comply with this rule, the range estimate which results in the smallest absolute difference

The target location is analytically calculated as the intersection of the ellipses defined by the associated range estimates. The location estimates with respect to each of the six sensor nodes is shown in Fig. 59(a), where the estimates are represented by the same color as the respective

**Sensor 1**

**Sensor 5**

**Figure 59.** a)The target position estimates by each sensor and b) the target tracks after data fusion

The location estimates from all sensors are fused together resulting in a single target location per target. The estimated target locations are not in a track form and contain a significant

Y [m]

**Sensor 4**

−7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8

X [m]

(b)

**Sensor 3 Sensor 2**

**Sensor 6**

**Sensor 1**

**Sensor 5**

−1

1

0

1

0

−7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7 8

X [m]

(a)

**Sensor 3 Sensor 2**

**Sensor 6**

ranges (black) are shown

is chosen.

sensor node.

Y [m]

**Sensor 4**

*9.2.3. Data fusion and target tracking*

0.5

0.5

0

1

The CoLOR project was devoted to the recognition of unknown environments using UWB technology. This topic encompasses a number of partial challenges. In order to obtain a complex picture of some catastrophic scenario, like the detection of victims after some natural disasters, their location in collapsed buildings, the geometrical information and the status of the buildings, we derived new detection, localization and imaging algorithms. Their performance was analyzed on simulated data and data measured in realistic scenarios using UWB sensors. These UWB sensors are capable of real-time operation in MIMO configuration. This allows us to analyze the application of UWB sensor networks and cooperative approaches for the localization of sensor nodes within the network, for the localization of people, for the detection and monitoring of their live signs, and for the imaging of their surroundings.

It was shown that by using a mobile UWB radar with multiple antennas, it is possible in an efficient way to reconstruct the basic layout of rooms and the position of freestanding objects. The detected features are added to a map while at the same time the own position is estimated (SLAM). To minimize the computational cost and the number of measurements needed, simplified models for wave propagation and stochastic, dynamic state space estimators where enhanced. The method of data association proved to be most critical regarding the precision and reliability of the map.

Using this map as a-priori information, the detection, localization and the imaging of the objects within an indoor scenario can be performed using the developed localization and imaging algorithms. By knowing the location of the individual objects, the potential of UWB radar was fully tapped by obtaining super-resoluted local information about 2D as well as 3D complex objects (concerning the outer contour). The interior of objects was gathered by novel algorithms which are based on exact radiation patterns depending on the permittivity of the medium while showing low computational load. The obtained radar images are post-processed by means of object recognition algorithms designed for full, fragmented or restricted illumination to provide recognition of the object under test from a finite alphabet. By the adaption of classical ellipsometry to the UWB-range, an estimation of dielectric surface properties can robustly be performed even for small dimensioned objects with a size of a couple of wavelengths. In addition polarimetric measurements as well as polarimetric data processing were taken into account to obtain object features which may remain invisible in mono polarized systems.

In order to test and compare different algorithms and antenna arrangements for indoor UWB sensing and imaging, a realistic UWB multi-path propagation simulation tool was developed. The propagation model is based on a hybrid approach which combines the deterministic ray tracing method (based on geometrical optics and the uniform theory of diffraction)

58 Will-be-set-by-IN-TECH 236 Ultra-Wideband Radio Technologies for Communications, Localization and Sensor Applications Cooperative Localization and Object Recognition in Autonomous UWB Sensor Networks <sup>59</sup>

with statistically distributed scatterers. Verification measurements show that the new model delivers very realistic channel parameters like channel impulse response, azimuth spectra, and path loss. Thus, it is suitable for an application in UWB system simulations.

[8] Cramer, R.-M., Scholtz, R. & Win, M. [2002]. Evaluation of an ultra-wide-band propagation channel, *IEEE Transactions on Antennas and Propagation* 50(5): 561–570. [9] Daniels, D. J. (ed.) [2004]. *Ground penetrating radar (2nd Edition)*, Vol. 15 of *IEE radar, sonar, navigation, and avionics series*, Institution of Electrical Engineers, London, UK. [10] Dardari, D., Chong, C.-C. & Win, M. Z. [2006]. Analysis of threshold-based TOA estimator in UWB channels, *Proc. Eur. Signal Process. Conf. (EUSIPCO)*, Florence, Italy. [11] Dardari, D., Chong, C.-C. & Win, M. Z. [2008]. Threshold-based time-of-arrival estimators in UWB dense multipath channels, *IEEE Transactions on Communications*

Cooperative Localization and Object Recognition in Autonomous UWB Sensor Networks 237

[12] Dardari, D. & Win, M. Z. [2006]. Threshold-based time-of-arrival estimators in UWB dense multipath channels, *Proc. IEEE International Conference on Communications*, Vol. 10,

[13] Deissler, T., Salman, R., Schultze, T., Thielecke, J. & Willms, I. [2010]. UWB radar object recognition for slam, *Proceedings of the 11th International Radar Symposium*, pp. 378–381. [14] des Noes, M. & Denis, B. [2012]. Benefits from cooperation in simultaneous anchor-less tracking and room mapping based on impulse radio - ultra wideband devices, *19th*

[16] Elsherbini, A., Zhang, C., Lin, S., Kuhn, M., Kamel, A., Fathy, A. & Elhennawy, H. [2007]. UWB antipodal vivaldi antennas with protruded dielectric rods for higher gain, symmetric patterns and minimal phase center variations, *Antennas and Propagation*

[17] Falsi, C., Dardari, D., Mucchi, L. & Win, M. Z. [2006]. Time of arrival estimation for UWB localizers in realistic environments, *EURASIP Journal on Applied Signal Processing*

[18] Fügen, T., Maurer, J., Kayser, T. & Wiesbeck, W. [2006]. Capability of 3-D ray tracing for defining parameter sets for the specification of future mobile communications systems,

[19] Gezici, S. [2008]. A survey on wireless position estimation, *Wireless Personal*

[20] Gezici, S. & Poor, H. [2009]. Position estimation via ultra-wide-band signals, *Proceedings*

[21] Gezici, S., Tian, Z., Giannakis, G., Kobayashi, H., Molisch, A., Poor, H. & Sahinoglu, Z. [2005]. Localization via ultra-wideband radios: a look at positioning aspects for future

[22] Guvenc, I. & Sahinoglu, Z. [2005]. Threshold-based TOA estimation for impulse radio UWB systems, *Proc. IEEE International Conference on Ultra-Wideband ICU 2005*,

[23] Janson, M., Fügen, T., Zwick, T. & Wiesbeck, W. [2009]. Directional channel model for ultra-wideband indoor applications, *IEEE International Conference on Ultra-Wideband,*

[24] Janson, M., P., J., Zwick, T. & Wiesbeck, W. [2010]. Directional hybrid channel model for ultra-wideband MIMO systems, *4th European Conference on Antennas and Propagation,*

*International Conference on Systems, Signals and Image Processing, IWSSIP 2012*. [15] Dutta, P., Arora, A. & Bibyk, S. [2006]. Towards radar-enabled sensor networks, *Information Processing in Sensor Networks, 2006. IPSN 2006. The Fifth International*

*Society International Symposium, 2007 IEEE*, pp. 1973 –1976.

URL: *http://www.hindawi.com/journals/asp/2006/032082.abs.html*

*IEEE Transactions on Antennas and Propagation* 54(11): 3125–3137.

URL: *http://www.springerlink.com/content/w1481x761266m417*

56(8): 1366–1378.

pp. 4723–4728.

2006: 1–13.

pp. 420–425.

*EuCAP'10*.

*Conference on*, pp. 467 –474.

*Communications* 44: 263–282.

*of the IEEE* 97(2): 386 –403.

sensor networks, 22(4): 70 – 84.

*ICUWB'09*, Vancouver, Canada.

Further within CoLOR a flexible and dual-polarized UWB antenna array has been developed. The major challenges beside the huge bandwidth itself were to design antenna elements which are able to meet the requirements regarding size, pattern, beam-width, polarization and the location of the phase center (over frequency). Via switches it is possible to select and control the single elements of this array, which allows its adaption to the different localization, imaging and object recognition algorithms and applications in this project. Due to the dual polarized antenna elements, the possibility to take advantage of polarization diversity is given and demonstrated.

Our results show that the UWB technology and especially the cooperative approach that fuses diverse information from multiple sensors provide a big potential for safety, security and emergency applications.
