Indoor Personnel Positioning And Walking Posture Recognition Based On Vibration Characteristics

In order to meet the increasing personalized needs of users in daily life,personnel positioning and walking posture recognition in indoor environment have been widely studied.Indoor positioning and walking posture recognition based on camera and wireless network cannot protect the privacy of users,and the equipment for postprocessing data has high requirements,so it cannot meet the needs of users in many cases.In the algorithm of human positioning and walking posture recognition,efficiency and accuracy also need to be improved.In view of the current demand,this paper selected the non-invasive acceleration sensor as the main equipment for monitoring indoor personnel,and proposed a scheme for indoor personnel positioning and walking posture recognition based on vibration characteristics according to the perception principle of scorpion positioning prey.First of all,this paper proposes a method based on non-invasive acceleration sensor for users' needs of privacy protection and mobility convenience.The Basitarsus Composite Slits Sensilla at the end of the leg of a scorpion can locate the strain caused by the prey vibration acting on the tarsal tip with an amplitude of 1nm.Inspired by scorpion positioning prey,the acceleration sensor was designed according to the position of scorpion BCSS,and a set of data acquisition platform imitating scorpion positioning biological structure was built.And the platform is used to collect indoor positioning data and walking attitude data of users.Through performance test experiments,the platform is verified to be stable and practical in function.Secondly,the higher the positioning accuracy,the better the positioning algorithm needs.This paper proposes a location fingerprint indoor positioning method that imitates scorpion positioning mechanism.The Leaky Integrate-and-Fire neuron model was used to simulate scorpion neurons,and the command neurons of scorpion under the combined action of excitation and inhibition were used to pulse code the vibration signal.Different input signals would generate different pulses,and the pulses were extracted as the location characteristics.This feature is more fine-grained than Received Signal Strength to characterize the characteristics of the Signal,so that the Signal has better discrimination.This method is applied to real location scene,which verifies that the location fingerprint indoor location method proposed based on scorpion location theory is more accurate than the location fingerprint location method based on RSS.Finally,based on the biological structure data acquisition platform of scorpion positioning function,a set of indoor human walking posture recognition system is built.The signal processing process of the system is as follows: firstly,the collected walking attitude data is preprocessed by detrending,wavelet denoising,etc.,and the time domain,frequency domain,time-frequency domain and pulse features of the data are extracted,and the features are normalized.Then the Hidden Markov Model was established according to the characteristics of human walking posture.A fast and automatic method for searching the peak data density was applied to establish the relationship between the internal parameters of the Model and walking posture characteristics.Then,Baum-Welch algorithm was used to train a hidden markov model for each walking posture.Finally,Viterbi algorithm was used to classify the walking attitude data,and the classification result with the highest similarity was determined as the classification result.The experimental results show that the system can effectively recognize walking postures such as normal walking,fast walking,running and stepping.