Using Sensor Equipped Smartphones to Localize WiFi Access Points

Determining the position of mobile devices is an important business and research area due to the many applications it enables. Outdoors, positioning is often solved by the Global Positioning System (GPS). Indoors however, the signals received from GPS satellites are too weak to provide meaningful position estimates. Therefore, indoor positioning must be done using other techniques, such as trilaterion based on WiFi signal strengths. The trilaterion technique requires that the positions are known for the WiFi Access Points (APs) used. While determining AP positions can be done manually, this is often time consuming, error prone and fragile to changes of AP positions. A better way is to use an automatic system for determining the AP positions. There are two contributions presented in this thesis. The first contribution is a navigation system using accelerometers and gyroscopes in a phone to track the movement of a walking person. A requirement of the navigation system is that the start position and heading of the user is known, and that the orientation of the phone does not change in respect to the user while navigating. The navigation system typically gives position estimates accurate to within a few meters for up to two minutes after navigation started.The second contribution is an algorithm that given a set of Received Signal Strength Indications (RSSIs) at different locations provides an estimate of where the transmitter is located. The accuracy of the algorithm varies depending on the coverage of measurements as well as the radio environment around the AP. The position of APs can typically be estimated within five meters from their true location when provided with measurements from several angles from the AP