Study On IMU/TOA Fusion-based Positioning Model And Performance Optimization For Human Motion Tracking

With the rapid development of the Internet of Things,wearable human motion tracking technology has been drawing more and more attentions,which has been widely used in many fields,such as medical care,emergency rescue and audio-visual entertainment.However,at present,wearable human motion tracking system mainly relies on inertial sensors,e.g.,accelerometers and gyroscopes.Its biggest challenge is that the error increases with time,which cannot be restored automatically and accumulate gradually.Namely,there exists accumulative errors and drift problems.Aiming to solve the accumulative errors and drift problems faced by traditional inertial wearable devices,this paper conducts performance evaluation and algorithm research based on IMU/TOA fusion-based positioning systems.The main contents and innovation points of this paper are detailed as follows:(1)Performance evaluation method of human motion tracking:By the derivation and analysis of theoretical error lower bound,it is proved that IMU/TOA fusion method is feasible to solve the problem of accumulative errors and drift of inertial sensors.With the derivation of Cramer-Rao Lower Bound(CRLB)of the fusion positioning system,an in-depth analysis was made from the perspectives of both spatial performance and temporal performance.The effectiveness of the IMU/TOA fusion method in solving the cumulative error problem was theoretically proved.Simulation results show that the fusion of IMU/TOA positioning system has large improvement in both spatial and temporal performance.It can provide theoretical basis for algorithm or system design,and work as an effective evaluation method for algorithm performance.(2)Geometrical kinematic modeling of human motion:By geometrical kinematic modeling,multi-source integration of IMU/TOA was achieved,and the problem of information fusion method for IMU and TOA was solved.This paper put forward a model of human motion tracking based on geometrical kinematics.The human body was considered as a whole connected by joints.Denavit-Hartenberg(D-H)equation was adopted to fuse the inertial measurement data and TOA ranging data,which to a certain extent,limited the accumulative errors and the drift problem.(3)Human motion tracking based on error estimation and optimization: The combined errors of IMU/TOA and optimization method were considered,to solve the problem of how to improve human motion tracking accuracy.Achebyshev center-based minmax optimization method was presented,which takes the real position of the target into account geometrically,in order to improve the tracking accuracy.By comparing with the state-of-the-art optimal target tracking methods,proposed method shows better measurement accuracy and consistency in different scenarios(that is,measurement variance).(4)Wearable platform based on IMU/TOA fusion:The aforementioned fusion human motion tracking model and algorithms were applied to the actual system design.A wearable measurement platform based on IMU/TOA fusion method was presented.The model and algorithms proposed in this paper were validated by the motion tracking experiment in actual scenes,which is of great practical application values.