Font Size: a A A

Research On Acoustic Direction Finding Using Smartphones

Posted on:2019-03-08Degree:DoctorType:Dissertation
Country:ChinaCandidate:X P WangFull Text:PDF
GTID:1318330542997985Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the development of mobile internet,indoor localization technology has widely application in our daily life.Due to the popularization of smartphones,the location based service has added to a large number of mobile clients,such as the Meituan and IKEA app,etc.The indoor localization technology has tremendous research value and research difficulties.Indoor localization project turns out to be a big challenge to the researchers.However,most indoor localization techniques require either specialized devices or extensive preparations(finger prints or floor plans).These techniques are inconvenience for daily use.To provide a novel solution which is high-efficiency,high accuracy and light weighted becomes a goal for many localization systems.According to the principle of antenna array for direction finding,we propose a novel scheme to find the direction of target anchor points(common speakers)using smartphones.Then we design and deploy our virtual antenna array direction finding system in various indoor environment.Specifically,there are three main aspects of our research:1.The structure of the Virtual Uniform Linear Antenna Array(VULA)For the high direction finding accuracy of an antenna array,and to make sure our direction finding system is easy to use in daily life,we propose a study on Virtual Uniform Linear Antenna Array.Our method is derived from a key insight:by moving a smartphone in a linear pattern,we can effectively emulate the sensitivity and functionality of the Uniform Linear Antenna Array.While we are moving the smartphone,the smartphone receives acoustic signals at successive locations in short distance.At each point in time,we captures a single measurement from the whole sampling data as if we had a receive antenna array at each of these sampling points.2.The structure of the Virtual Uniform Circular Antenna Array(VUCA)Due to the azimuth range limitation of the Uniform Linear Antenna Array,it only provides[-90°,90°]azimuth estimation.Anyway the direction finding system need to adapt to any specific scenarios in indoor environment.Thus we provide a direction finding algorithm based on Virtual Uniform Circular Antenna Array.Instead of shaking or moving the smartphone in front of users' body,a user can hold the smartphone still and rotate his body around for 360 degrees during sam-pling to imitate a Virtual Uniform Circular Array.The Virtual Uniform Circular Array can provide omnidirectional direction finding and it will also estimate the elevation angle of the target anchor point.3.Virtual Uniform Antenna Array direction finding systemWe designed,deployed,and evaluated our scheme of direction finding under diverse circumstances.We preset some regular speakers transmitting high fre-quency acoustic signals in the indoor environment as anchor points.Then after signal processing and signal modeling,we leverage multiple signal classification to calculate the direcition of arrived signals.Due to the complex indoor environ-ment,indoor direction finding system would easily be affected by the multipath signals.The key to reduce the multipath effect is leveraging multiple signal clas-sification algorithm for DOA calculating.Besides,for different user motion pat-terns,we also leverage the internal sensors of a smartphone to correct the errors caused by non-uniform virtual antenna array.Imitating a virtual antenna array by ordinary smartphone and overcoming the com-plex indoor scenarios are the major challenges for implementing our system in indoor environment.The fickle sampling motion patterns of the smartphones also have side-effect on our direction finding system.After addressing these challenges with some novel algorithms and deploying our system under various circumstances,the extensive comparative experiments show that our system is efficient and accurate.
Keywords/Search Tags:indoor localization, indoor direction finding, antenna array simulation, multiple signal classification, multipath effect
PDF Full Text Request
Related items