Remarkable advances have been made in the development and technical optimisation when considering stationary solutions of assistant systems for the visually impaired. For example, recent progress in web accessibility allows sensory handicapped people to receive in many cases the same amount of information as sighted persons. While looking at mobile solutions, there are still many other challenges to achieve equal opportunities concerning information access. In 2004 we presented a design for a new type of an indoor navigation and object identification system for the blind [1]. The basic idea of this system is to combine local sensor information with 3D environment models. The local sensor information can be acquired using a hand-guided sensor module. This sensor module consists of a stereo camera, a 3D direction sensor, and a keyboard. It is also possible to attach the sensor module to the cane. By pressing keys, inquiries can be sent either to the connected portable computer or to a platform of one or more servers distributing information about the closer and distant environment. The location of the user can be determined using conventional WiFi installations. The system enables users to recognise modelled objects, i.e. the closest object in front of the user. The name of this object is transmitted to the user over a text-to-speech engine.