Research On Key Techniques Of Motion Monitoring System Based On Bluetooth Low Energy

Wearable movement monitoring system is the current research focus, technology companies and foreign scholars have done a lot of work in this field, and has launched a number of well-known products. However, today’s research and existing products, also have a lot of defects in signal processing, motion state identification and other aspects, accuracy is also uneven, and research methods are relatively limited. Marketing aspects of the products is much larger than technical factors. Therefore, the research field is still in the exploratory stage, and has no fully mature solutions.In this paper,we regard sensor subsystem of motion monitoring system as the research object. We object low-power Bluetooth(BLE) as means of communication, and design a sensing device with a triaxial accelerometer. According to data analysis and actual conditions, we made an intensive research of algorithms, methods, hardware and software design and other aspects involved in the implementation of movement monitoring. The main findings include the following aspects:Firstly, combining the simulation results, this paper presents an implementation method of digital filter applied on wearable devices. According to the simulation results, we designed the digital filter for motion monitoring. After filtered, Smooth signal waveform can be obtained, and eliminate the influence of signal burr.Secondly, this paper proposes a multi-threshold- peak meter step algorithm, and establish a model of motion state machine. According to cyclical variations of acceleration in trunk movement and arm swing, we extract characteristic values for step detection and gait recognition, and excluding the impact of external acceleration disturbance. Then, the simulation Matlab program is built to validate the algorithm. The state machine uses a nested structure, divide into two layers, states and their substates. And the transition conditions are determined.Thirdly, the paper proposes a verification method on effectiveness of feature parameters based on SVM classification algorithm. And we use the feature parameters selected in this design as example to introduce the method. Then, through the analysis of simulation results, the effectiveness and sensibility of the feature parameters in the design are discussed. In addition, this method is also analyzed by the classification prediction results.Fourthly, complete the implementation of motion detection applications. On application layer of BLE protocol stack, combined with the operating mechanism of the operating system abstraction layer(OSAL), the programs of meter step algorithm and state machine are compiled on the sensor device, which use the uses the means of step execution and state transition.Fifthly, complete the design of application profile of motion monitoring. Definite the motion monitoring service in BLE protocol stack, and standardize the service characteristics to ensure unified understanding of motion monitoring service between sensor devices and intelligent terminals.Sixthly, the comparison tests between the proposed motion monitoring system and other two well-known smart wristbands show that the proposed system in this paper has a good performance in motion monitoring.