In-Network Evaluation of Body Carried Mobile Terminal Performance

A way to characterize the overall performance of a mobile handset in the operating network is to collect the information transferred on the Abis during calls and process the data to obtain, for example, the average receive power level. In this way the actual GSM network becomes the measurement system and the performance evaluation can be done when the handset is in normal use. In co-operation with the Danish GSM network operator TeleDanmark some measurements were made where four different commercially available handsets were tested, labelled handset A-D. The measurements were made in a number of ways in two different environments. In one series of measurements 18 test users carried the handsets in two different positions on the body, in a chest pocket and on the hip [Nielsen, Pedersen, 2001b]. This was done in order to investigate the body loss observed when the user is not holding the handset in the hand and next to the head, but for example is using a hands-free kit. Also a series of measurements involving a full body phantom was carried out. CPK Annual Report 2001 - Figure 4.5.8b (Mean link loss for body carried handsets) shows an example of the losses in the channel observed with different users. The mean body loss was found to vary between about 3.5dB and 9.5dB, mainly depending on the environment and the position of the handset on the body. For the same environment and handset, the difference in body loss observed for the two positions on the body was typically 1-2dB. The difference between the two environments was typically 2.5-3dB. The body loss observed for the different types of handsets in the same conditions varied up to about 1.5dB. Handset size itself is not a distinguishing parameter, and handsets with internal antennas are not generally inferior to handsets with external antennas. However, it does seem that the handsets with internal antennas are more sensitive to position on the body. The handset performance variations among the users have a standard deviation of typically 1-2dB, but peak differences up to 9dB have been observed. The body loss for the measurements involving the full body phantom did not match what was found for the real user case. Both the absolute values of body loss as well as the relative body loss of the handsets measured with the phantom differs significantly from the mean values obtained with the test users. (Jesper Ø. Nielsen, Gert F. Pedersen)