| Nowadays,Mobile Crowd Sensing(MCS)develops rapidly with the popularity of Mobile intelligent devices.In MCS,a large number of users with mobile devices can collect large-scale,highquality sensory data anytime and anywhere,and through the rational use of these data to meet the sensory needs of data demanders and help the construction of a modern digital city.Currently,task assignment as a key research direction in MCS has been widely studied by domestic and foreign experts and scholars.At present,most of the studies have been conducted on general type tasks at specific time and specific locations,and few consider the time-critical task assignment problem where the area scale is restricted but the specific location is not.For this type of tasks,the key issues to be solved are how to help task publishers set the data collection area and how to select the task participants who respond timely to perform the tasks.(1)To help task publishers determine task collection areas,a Clustering In Hot Areas(CIHA)algorithm is proposed.First,mobile vehicles are selected as task participants,the map area is rasterized,each grid is defined as a sensing interest area,and the sequence of moving trajectories of vehicles is simplified using the sequence of perceptual interest regions.Then,with the goal of helping task publishers to identify task collection areas,the algorithm gradually discovers hot areas based on the density distribution of vehicles in the grid and uses them as task collection areas.Finally,the feasibility of the CIHA algorithm is verified with a real dataset.(2)A Hidden Markov Model(HMM)-based Next Hot Area Prediction(NHAP)and a Task Assignment Based on Response Time Optimization(TABRT)algorithm are proposed to enable task participants to respond to tasks in a timely manner,respectively.The former fully considers the characteristic of vehicle mobility and uses the HMM model to establish the association between the sensing interest areas and the hot areas,so as to predict the hot areas where the vehicle will be located at the next moment.The latter takes into account the distance between the vehicle and the perception task in the hot areas as well as the historical speed of the vehicle based on the vehicle prediction,and uses a greedy algorithm to assign tasks to the vehicle enough to optimize the vehicle response time to the perception task to the maximum extent.The experimental results show that NHAP has good performance in vehicle prediction accuracy,while the TABRT algorithm significantly increases the completion rate of task assignment as well as reduces the vehicle response time to the task.(3)Based on the above algorithm,a prototype MCS task assignment system was designed and developed,and the Web side and mobile side were designed according to different types of users.The platform administrator mainly manages the sensing tasks through the Web side,while the task participants respond to the sensing tasks through the mobile side.In summary,the CIHA clustering algorithm proposed in this thesis can effectively delineate hotspot areas as well as the TABRT task assignment algorithm can effectively optimize the response time of time-critical task assignment,which has certain theoretical significance;the developed prototype system has certain application value after being adapted and deployed by relevant practitioners with actual application scenarios. |
