Research On Indoor Location Based On Smart Phone And PDR

With the rapid development of location services and mobile Internet,the demand for indoor positioning and navigation is increasingly urgent.In the indoor environment,the global satellite navigation system signal is attenuated,so GPS cannot provide accurate positioning information.Considering multiple factors such as positioning accuracy and construction cost,indoor positioning method based on micro inertial sensor with smart phone as the medium is paid attention.Based on pedestrian dead reckoning positioning technology,this paper studies how to use the inertial sensor integrated in smart phones to accurately detect gait,estimate step length and estimate heading.The specific research contents and innovations of this paper are as follows:1.A gait detection method based on peak-valley pair constraint is proposed.For the traditional peak detection method with only threshold constraint,it is unable to eliminate the problem of excessive step counting caused by false peak value.In this paper,pedestrian gait is analyzed based on the theory of kinesiology,and the relationship between motion acceleration and pedestrian gait cycle during walking is obtained.on the threshold constraint basis of traditional peak detection method,peak and valley in acceleration is put forward to correspond in pairs to be recorded as effective peak-valley value,and then get the number of walking steps.Experiments show that the accuracy of this method is higher than that of the traditional peak detection method when the mobile phone is in handheld,swing and pocket mode.2.A modified step-length model based on Weinberg method is proposed.Aiming at the problem of frequent fluctuation of step size estimation curve by Weinberg method in three sets of asynchronous speed experiments,considering that the adjacent step size of people in normal continuous walking process has a certain continuity,the k-1st step size estimation result was introduced into the step size estimation result by Weinberg method,and the k th step size estimation was carried out by assigning different weight factors.Experimental results show that the improved step size estimation model has a slight improvement inaccuracy compared with the Weinberg method.3.A heading correction method based on corner matching and kalman filter is proposed.In this paper,the heading Angle data output by the direction sensor of the smart phone is mainly used,the heading Angle curve after unwinding is still jitter frequently in the straight phase.According to the gyroscope angular velocity,the rotation Angle is detected,the heading estimation is corrected by kalman filter in the stage of no turning,and the original direction sensor data is retained during the turning stage.This method maintains the trend of original heading Angle data and reduces the amplitude of data jitter,thus improving the accuracy of heading estimation.In conclusion,this paper studied gait detection based on peak-valley constraint,modified step-length model based on Weinberg method,and a heading correction method based on corner matching and kalman filter,and conducted experiments on MATLAB to verify the reliability and effectiveness of the above research work.