The aim of this diploma thesis is the development of an ad-hoc-application based on web services and Peer-To-Peer communication. An existing "Universal Speech Control" (USC) developed at ZHW is to be extended into an ad-hoc-application so that any device found by the speech control in the network can be remote controlled by voice. The existing USC builds a command tree from the description of the different devices to be controlled. Based on this tree, the speech commands are recognized and translated into logic commands for the devices. In the original speech control, this logic commands are only displayed but not sent to any devices.

The USC shall now be extended, so that any device can spontaneously contact the USC. The speech control again builds a command tree for all devices found. The speech commands shall then be recognized and the corresponding logic commands sent to the actual device.

When the control center, comprising the speech control and the device control, starts up, it look for all devices that can be controlled by voice. This is done by P2P-communication. The control center collects all information from these devices needed to control them by voice. After all devices have been found, the speech control builds its command tree and shows it to the user.

The devices provide web services for their control. In order to be found by the control center, the devices regularly publish their web services with advertisements in the network. For the creation and publishing of the web services the GLUE framework was used, for the P2P-communication the JXTA-framework from Sun.

The control center also generates its web services and publishes them, so that devices, which are connecting later, are able to find the control center and connect to it. The control center then updates its command tree to allow the control of the additional device.