Study On Multi-modal Data Fusion

Today, the use of sensors is becoming increasingly popular; it gives people an unprecedented level of convenience, and has a tremendous impact to the development of society. With the increasing number of sensor types, how to use the sensors more effectively to obtain more accurate information has become the focus of research. Therefore, multi-modal data fusion research has gained more and more attention.This thesis carried out in-depth study on the principle of multi-modal data fusion, multi-sensor fusion and the application of multi-modal data fusion in a vehicle collision avoidance system.Firstly, this thesis summarizes the development status of multi-modal data fusion technology and vehicle collision avoidance systems at home and abroad, and then discusses the existing problems.Then this thesis studies the sensors used by vehicle collision avoidance system, including speed sensors, acceleration sensors and a variety of distance measuring sensors, proposes a fusion algorithm of velocity and acceleration, and a fusion method of a variety of ranging sensor data. The simulation experiments are done in matlab environment.Finally, this thesis designs a vehicle collision avoidance warning system using the multi-modal data fusion method, and programs and tests it in the Eclipse environment.This thesis has done useful work in multi-modal data fusion.