Research On Indoor Positioning Method Based On Smartphone MARG Sensor

With the rapid development of mobile devices and wireless communication technology,location service has attracted more and more attention.In outdoor environments,the Global Navigation Satellite System(GNSS)can provide high-precision positioning services.However,in the indoor environment with coverage and lack of strong signals,GNSS signals will decay sharply and it is difficult to meet the requirements of positioning accuracy.Among many indoor positioning technologies,Pedestrian Dead Reckoning(PDR)has the advantage of high positioning accuracy in a short period of time,but it has the problem of error accumulation over time.At the same time,the geomagnetic signal is seamless and stable in the indoor environment.However,there is a problem of low resolution due to similar magnetic field characteristics in different positions.In order to achieve the complementary advantages of the two,this paper uses smart phone MARG(Magnetic Angular Rate and Gravity)sensor to collect data,and proposes a PDR and geomagnetic fusion location algorithm based on improved particle filter.The specific research contents are as follows:1.Aiming at the defects that the attitude calculation is susceptible to magnetic interference in PDR,the complementary filtering algorithm is improved in this paper.In the improved algorithm,the magnetic interference is detected by setting the threshold value of magnetic field modulus and magnetic inclination angle,and the magnetic interference is suppressed by dynamically modifying the filtering parameters to improve the accuracy of attitude calculation.2.In order to solve the problem that the single step length estimation model in PDR is difficult to accurately predict the step length in complex walking state,this paper proposes an adaptive hybrid step length estimation model.The model integrates linear and nonlinear models,makes full use of the acceleration and step frequency information when pedestrians are moving,and then accurately estimates the step length.At the same time,it was found in the experiment that there is a significant difference in the acceleration variance of pedestrians under normal walking and running states.Therefore,in this article,we distinguish between walking and running states by setting the acceleration variance threshold,and then dynamically adjust the step length model parameters to achieve the purpose of accurately predicting the step length under complex gait.3.Aiming at the problem of location drift caused by the accumulative error of PDR and the low resolution of indoor magnetic fingerprint features,this paper uses the improved particle filter to fuse magnetic field information to modify the PDR motion model,and finally achieve the effect of complementary advantages of the two positioning methods.4.Aiming at the problem of weight degradation and particle diversity depletion in the particle filter algorithm,and at the same time to ensure the efficiency of the algorithm,this paper introduces the firefly algorithm to optimize the particle filter.By using the global optimal value in the particle swarm to guide the remaining particles to update their positions,so the particles tend to move to the high-likelihood region,which can more accurately describe the true state of the target observation and improve the global optimization ability of the algorithm.At the same time,in order to avoid the particles falling into the local optimum and repeatedly oscillating at the extreme point,this paper optimizes the optimization method of the firefly,redesigns the brightness formula,attraction formula,step adjustment factor and position update formula.Finally,it was named Adaptive Optimization Firefly Algorithm(AOFA).In order to verify the effectiveness and stability of the proposed algorithm,this paper conducted many experiments in the indoor environment.The results show that the accuracy of the step detection algorithm in this paper can reach 95.5% under the conditions of forward,backward,running,up and down stairs,etc.In different phone holding states,the accuracy rate can also reach more than 91%.In addition,the average errors of the step length estimation model proposed in this paper are all less than 3.5 cm under different motion modes,the average error of attitude calculation is also within 4°.Finally,the overall positioning scheme of this article has been repeatedly tested under straight,turning and rectangular routes.The test results show that the positioning scheme of this article can provide positioning accuracy with an average error within 0.5 m,which can fully meet the indoor positioning requirements and has high application value.