A Target Recognition Mehtod Based On The Doppler Effect In Mobile Networks

Recent years, computer technology develops very quickly. And nearly everyone has a mobile phone. It has become a widespread phenomenon now which led to a lot of new applications, such as interaction between mobile devices in their close proximity. And the premise is that we need select the target device in close range. This thesis propose the method of the target device recognition based on Doppler Effect with high precision of recognition and without any additional configuration.This recognition process is completed by a gesture that the user point his device toward the target device. The idea is that we let the target device performs different character from other nearby devices. And if the character is obvious and the method of character testing is accurate, the select accuracy of the target device will be more accurate. Accordingly, this thesis uses the technique based on the Doppler Effect. And we can compare the calculated doppler shift to select the maximum doppler shift, then we can select the target device. The current mobile devices all have microphones and speakers, so the use of acoustic technique will not limit the scope of the application. In order to improve the accuracy of selecting the target device, this thesis designs and implements a band-pass signal processing pipeline to against ambient and acoustic noises. And this thesis uses the FFT algorithm to estimate the frequency shift of each receiving device. Then the Cross-Ambiguity-Function is proposed to improve the estimate accuracy of the frequency shift. At last, we can select the target device by searching the maximum frequency shift.Based on the method of the target device recognition described above, this thesis implements a system to select the target device and tests the system under different environment. The results of the tests show that this system can provide about 95% device selection accuracy within 3 meters for most of the scenarios. And it can provide about 90% device selection accuracy within 4 meters. When the distance is 5 meters and the angle between the source device and the receiving devices is not less than 30 degrees, the accuracy of the selected device can still reach over 85%.