Research On Energy Detection Of Energy Consumption Of Human Body Based On Inertial Units

As the core technology of smart medical care,smart sports,smart health and other fields,the research on energy consumption of human activities has become increasingly prominent with breakthroughs in these fields.Since the energy consumption of the human body in daily activities is closely related to various factors inside and outside the human body,it is very difficult to accurately calculate the energy consumption of the human body,which has also become the bottleneck of the research on human energy consumption.At the same time,the impact of different types of physical activity and exercise on energy consumption is different,but the current existing detection methods of human activity energy consumption are not enough to solve the problem of accurately calculating human activity energy consumption in the change of activity type in daily life.Therefore,there is an urgent need for an algorithm with high accuracy of human activity energy consumption under different activity types.This paper mainly studies the energy consumption detection technology of human daily activities based on inertial unit.It includes three research contents,human activity recognition algorithm,human activity energy consumption research,and human activity energy consumption detection system implementation,as follows:Aiming at the difficulty of modeling complex time series data in human activity recognition algorithms,this paper proposes a CNNBiLSTM algorithm based on convolutional neural network and long and short-term time series network.After mining data features through convolutional neural networks(CNN),they are used as input to bidirectional cyclic long-term mermory neural network models(BiLSTM)for human activity recognition.Secondly,this paper uses the features extracted by the convolutional neural network and the final output activity type as the input of the human activity energy consumption model,which provides strong support for accurately calculating the human activity energy consumption.In view of the study of human activity energy consumption,the influence of different types of physical activity on energy consumption is different,and the error in the calculation of energy consumption is large.In this paper,a human activity energy consumption model based on genetic algorithm and feature description is proposed to optimize artificial neural network.Firstly,in order to accurately calculate human energy consumption,the type of human activity is introduced as a characteristic variable.Then,the feature variables of the input artificial neural network are optimized,and the influence of the feature variables on energy consumption is strengthened through their quantitative description.Finally,by combining genetic algorithm with artificial neural network,the problem that the network only obtains the optimal solution locally is solved,and the accuracy of human activity energy consumption detection is improved.Finally,this paper designs and implements a human activity energy consumption detection system.Aiming at the application problem of human activity energy consumption research based on inertial unit,through the analysis of the existing human activity energy consumption detection system,this paper designs and implements the human activity energy consumption detection system based on the proposed CNN-BiLSTM human activity recognition algorithm and the human activity energy consumption model based on genetic algorithm and feature description optimization artificial neural network.Verified the expected functionality and performance of the system.The practical application of human activity energy consumption detection system has been realized.