Research On Energy-Efficiency Control Strategy Based On Context Awareness In Wearable Computing

With the rapid development of micro controller unit(Micro Control Unit,MCU),micro-electro mechanical system(Micro-electromechanical Systems,MEMS)and wireless sensor network(Wireless Sensor Network,WSN),the activity recognition system based on wearable technique and wireless communication technique has great application prospects and commercial value.Inertial measurement unit(Inertial Measurement Unit,IMU)has many advantages,including small volume,light weight,portable,low-cost,and extensive application in the field of medical rehabilitation,motion capture,and so on.However,long-running and real-time stream transmitting at high frequency may lead to the shorter battery life and poorly energy efficiency.To combat them,this thesis explores the strategies of improving the energy efficiency of wearable computing for motion tracking,which combined the technology of situational awareness.It improves the detecting accuracy,energy efficiency,and easily to wear.The energy efficiency is analyzed on Shimmer2R-wearable platform.The main research works follows:(1)Proposing a wrist-worn system,using single 3-axis accelerometer for human motion tracking.Currently the human movement recognition system based on 3-axis accelerometer has the problems such as cable containment,more sensor nodes to wear,heavy wearing equipment,inconvenient carrying,and so on.To combat them,this thesis proposes a wrist-worn system,using single 3-axis accelerometer worn on the right wrist of the experimenter for data acquisition.On the premise of not affect users' daily activities,it can meet the comfort of wear and can reduce the computational complexity.(2)Implementing a parameter optimization algorithm suitable for classification of human motion recognition.The algorithm improves the overall energy efficiency of the motion recognition system for human motion tracking.Because the former classification model usually adopts the default parameters,which cannot deal with the behavior characteristics of different individuals.In this thesis,the parameters of different classification models is optimized.And the best combination of parameters is calculated based on different individuals' exercise habit,so as to improve the classification accuracy of human motion recognition system.Experimental results show that the optimization algorithm can improve effectively the classification accuracy of the human recognition system.The classification accuracy can increased by 26.05 percentage points,and the highest classification accuracy can reach 98.59%.(3)Proposing an adaptive control algorithm and multi-classifier fusion algorithm for wireless transmission.The energy consumption of the wireless wearable node used in the motion capture system mainly concentrates on the wireless transmission communication.Real-time high-speed wireless transmission consumes most of the sensor node energy.In this thesis,adaptive wireless transmission control algorithm is proposed,combined with situational awareness technology.It can effectively reduce the energy consumption of wireless transmission by judging the processing of human motion data and adaptively controlling the data amount of wireless transmission.Experimental results show that the proposed adaptive data transmission control algorithm can reduce the amount of data transmission by 76.7%.By this method,the energy efficiency of the equipment can be improved and the lifetime of wireless sensor nodes can be prolonged.The maximum endurance of the node can be extended by 73.38%.