Research On Key Technologies Of Mobile Crowd Sensing Based On Collaborative Computing

Since the 21st century,technology has ushered in rapid development.The emergence of smart wearable devices and smartphones has changed the way in which fixed sensors collect information in the past.Mobile crowd sensing provides new ideas for the development of the Internet of Things,relying on GPS,camera equipment,microphones and other sensors.It provides a variety of data types for the system,and has become a new way of data sensing through its mobile flexibility and low cost advantages.In the crowd sensing system,due to its mobility,compared with the traditional static sensing network,the coverage is wider and the layout is more flexible.Different users and devices can be selected to select tasks,and different schemes can be formulated for different needs.However,the mobile crowd sensing system collects data from multiple users,only from the perspective of platform tasks,ignoring the cooperation and interaction between users,which may lead to data redundancy and load imbalance between devices,thus affecting the quality of uploaded data,increasing the time of sensing tasks and reducing the performance of the sensing system.Therefore,this paper studies the collaborative computing technology in crowd sensing system.The main research results are as follows :1.An optimized user selection recommendation method based on sparrow search algorithm is proposed.Firstly,the perception user is modeled,and the concept of user fitness is proposed.The basic information of the user is classified into four aspects :location,power,equipment and reputation,and the fitness is calculated in turn.Secondly,according to these fitness values,the priority of the perceived user is considered comprehensively,and the user is classified according to the user priority.The intelligent optimization algorithm is used to simulate the process of the user completing the task.Finally,through the sparrow search algorithm iteration process,the optimal user suitable for the perceived task is selected.Through the comparative experiments of the proposed algorithm and other optimization algorithms in the same environment,the results show that this method can significantly improve task allocation and data quality at a lower budget.2.A collaborative device offloading method based on edge computing is proposed.Firstly,using the deep reinforcement learning framework,the offloading decision is divided into two parts : generation and update.Secondly,in the generation part,all the offloading strategies are obtained through the learning of neural network according to the information of equipment and computing power,channel transmission delay,etc.,and then the best offloading strategy is generated according to the specific environment through reinforcement learning.Finally,in the update part,the historical generation strategy and the transmission channel gain are combined to update and train the neural network.Through device collaboration,the operation of data offloading computation solves the sensing scenario with a large amount of data,reduces the occupancy rate of some sensing devices,and reduces the computation delay.3.A task collaborative subcontracting method based on Nash equilibrium is proposed.Firstly,a Markov chain based cognitive user mobility model is established to analyze the spatial distribution of users.Secondly,according to the spatial distribution of the data collected by the task,the utility function is designed to analyze the execution ability of the user to complete the sensing task.During the movement of the user carrying the task,it will encounter other users.Finally,by comparing the utility function values,the sensing task is decomposed and subcontracted to adjacent users,and the Nash equilibrium is used to comprehensively consider the reward values of the sender and the receiver.The scenario of complex sensing tasks is solved by task subcontracting offloading,which improves the data coverage and task completion rate of sensing tasks.