**4. Conclusion**

10 Will-be-set-by-IN-TECH

function of the tissue parameters. The first row shows the intersect point for effective antenna heights of a quarter wave length of the respective frequency. The second row shows the intersection point for a fixed effective antenna height which has been set to a quarter wave length of the lower UWB edge frequency (*f* = 3.1 GHz). The comparison of the results shows that with increasing frequency even at distances greater than 10 m the surface wave component is lower than the corresponding space wave. This fact implies a relatively weak far field and causes a reduction of the directivity at high frequencies. With respect to the design of future UWB on-body antennas this circumstance has to be considered. Additional investigations have also shown that vertical polarized antennas excite a much more dominant surface wave than equivalent horizontal orientated antenna configurations, see [7]. These results are in accordance to the theory given by [8] and should be considered to optimize

The validation of the suggested model with respect to the anatomical structure of the human body, with its numerous tissue types and curved surfaces, is done by a path gain calculation of a complete human body voxel model. Basis for this derivation is the numerical IT'IS virtual family Duke model [12]. The selected scenario consists of a transmitting antenna TX which is located at the right shoulder front. The corresponding receiver RX is shifted along the front side of the trunk above the right leg to the right foot. Figure 6 shows the path gain along the chosen path *d*. In addition, the path gain of the suggested on-body model has been calculated for homogeneous muscle and fat tissues. As seen in Figure 6, the calculated path gain of the

**Figure 6.** Path gain of the Duke voxel model in comparison to the homogeneous model of fat and muscle; Additional included is a numerical validation graph of a layered model analog to [7].

Analog to [7] a numerical model of a layered plane surface has been implemented to realize a more realistic representation of the human tissue structure. The suggested simulation model consists of 2 mm skin and 5 mm fat tissues which are positioned on an infinite muscle tissue.

UWB applications for given propagation scenarios.

voxel model lies between the graphs of the theoretical models.

*3.1.2. Limitations of the on-body model*

The study has shown that an antenna de-embedding for in- and on-body applications can be realized by the derivation of corresponding far field models with reasonable accuracy for practical applications. Related to this theory, quantities as the directivity and the effective antenna area have been defined to derive good approximations of propagation models. Especially for on-body applications the suggested model gives a detailed insight by the separation of the electromagnetic field in its space and surface wave components.

Moreover, the presented theory enables the calculation of average path gain models of arbitrary antennas which can be reduced to a source of vertical and horizontal orientated current distributions. By this, the numerical calculation space can be reduced to the minimum far field distance of the corresponding model. Additionally, an insulated UWB tear drop antenna design has been presented for in-body communication applications to give an adequate validation example. For the on-body scenario the UWB teardrop antenna has been modified and also been discussed.

In future studies the in-body approach has to be modified to a multipath channel model to include additional propagation effects like surface waves. In addition, the on-body model has to be extended to give a wider applicability with respects to the complex structure of the human body. Moreover, the effective antenna area for on-body applications has to be described as function of the given model. With these improvements a structured combination of the on-, off- and in-body scenarios seems realizable to develop an optimized antenna theory for body centric communications.
