Research On Indoor Positioning Method Of Mobile Car Based On Multi-sensor Fusion

In the outdoor environment,the technology of positioning by the GPS has been relatively well developed,But in the indoor environment,due to the obstruction of buildings and the complexity of the indoor environment,their location precision is poor,usually the location information is obtained through the sensors carried by the mobile robot itself.However,since the characteristics of each sensor are different and the applicable situations are also different,it is difficult to adapt to various complicated environments by simply relying on one sensor for positioning,and the accuracy of positioning cannot be guaranteed.The positioning method based on multi-sensor fusion has important research value as it enhances the stability of the system and improves the precision of positioning by integrating data from multiple sensors.In this paper,the location of mobile car in an indoor environment was studied.Main research efforts are as follows:1.First,a wheeled mobile car experimental platform was built.A four-wheel differentialdrive car was used as the main body of this platform,and the microcontroller STM32F407 was adopted as its core controller.Meanwhile,the mobile cart was equipped with a number of modules,including photoelectric encoder,ultrasonic module,laser radar and wireless communication module.Secondly,the programming for the microcontroller and the upper computer was finished.Finally,the wireless data transmission of the lower and upper computers and the basic control and positioning functions of the mobile trolley were realized.2.A fast location algorithm is proposed based on the speed information of the mobile car.Firstly,the relation between the true position and the position in the image is established,based on which,the position of the mobile car could be obtained directly from the distorted image.And then,the speed information is used to located the car in the image,which will reduce the searching time.3.Based on the extended kalman filter and the unscented kalman filter,location methods that integrate data from photoelectric encoders and laser radars were proposed.Firstly,the movement model of the mobile car is established? and then the measurement model of photoelectric encoders and laser radars were established.The multiple sensor-based location methods for were proposed based on the extended kalman filter and the unscented kalman filter respectively.Finally,the algorithms were verified by the experiment.