| With the popularization of intelligent mobile devices and the development of 5G communication technology,mobile crowd sensing as a new way of perception,has been more and more applied to actual scenes.It uses various sensors built in mobile terminals to complete the perception of the surrounding environment.In mobile crowd sensing,perception nodes are usually characterized by high mobility and dense distribution.Therefore,compared with the traditional fixed sensor network,mobile crowd sensing network has a wider perception scale,lower deployment cost,and better comprehensiveness and timeliness of perception data.However,the uncertainty of the perceived node increases the difficulty in acquiring the perceived data,such as the data collection ability of the perceived node,acceptance of the communication cost,movement path,personal willingness and preference,etc.In order to ensure the quality of perception data,it is very important to match appropriate perception nodes for data collection according to the content requirements and spatial and temporal distribution of perception tasks.Therefore,task assignment and perceptual performance evaluation in mobile crowd sensing are studied in this paper,and the main achievements are as follows:1.A task recommendation method based on user-hybrid model and collaborative ranking is proposed.First of all,according to the historical behavior of participants,the interest preference,movement characteristics and other aspects are analyzed to filter out low-quality sensing users initially.Meanwhile,the similarity among the participants is used to build a user-hybrid model.Then,the participants’ willingness will be predicted by the probabilistic matrix factorization,and a ranking model is obtained through the supervised sequential learning algorithm.Finally,a task recommendation list is generated on the basis of ranking model as the preferred task list for the target participants.As a result,the sensing tasks are recommended to the participants with the minimum movement distance and the highest acceptance rate.The simulation experiments based on the real dataset show that the method proposed in this paper can improve the accuracy of task assignment effectively and reduce the moving distance of sensing users simultaneously.2.A Two-phased participant selection method based on partial transfer learning is proposed.Firstly,the data is preprocessed.On the one hand,the sensing task features are extracted,so as to analyze the correlation between the source task and the target task feature space.On the other hand,users are divided into active users and passive users according to the historical movement law of sensing users.Secondly,the task assignment in the first stage is carried out.According to the similarity of feature space between the source task and the target task,migrate part of the user resources of the source task to a target subtask with a similar distribution of its feature space.So that the target task can efficiently and accurately select participants.Finally,the task assignment in the second stage is carried out.For the target subtask not covered,the passive user in the subtask area is taken as the assignment object.Simulation results based on real data sets show that this method can effectively improve the task coverage and reduce the perceived excitation cost.3.A sensing performance evaluation method based on performance evaluation process algebra is proposed.First,a multidimensional evaluation index system is constructed according to the sensing task demand and sensing user behavior.To determine a sensing performance evaluation standard covering a variety of requirements.Secondly,based on the Performance evaluation process algebra,a formalized model is established to accurately describe the behavior and properties of crowd sensing compute.Meanwhile,using Markov Chain to grammar semantics verification and model analysis,realize to perceive the quantitative evaluation of the sensing.Finally,according to the results of quantitative calculate the instantaneous incremental perceived performance.Based on incremental change,the trend of sensing performance is predicted,and the direction of sensing performance optimization is determined. |
