Research On Task Allocation Of Group Intelligence Perception Under The Constraint Of Ability Diversity

With the popularity of mobile smart devices(such as smart phones,tablets and car-sensing devices)and the rapid developments of sensing technology and communication technology,crowdsensing has attracted more and more attention.In crowdsensing,ordinary users use their mobile smart devices(such as smart phones,tablets)as basic sensing units,and conduct conscious or unconscious cooperation through mobile Internet(such as WiFi,cellular networks and wired networks),to achieve sensing tasks distribution and sensory data collection,which completes massive and complex sensing tasks.Compared with traditional sensing methods,crowdsensing has many the characteristics,such as flexible economical sensor deployment,unreliable multi-source data sensing,and wide coverage.By using the extensive participation of a large number of ordinary users,crowdsensing can complete complex social sensing tasks at a lower cost,with important research significance and application value.Task assignment is the basis of the application in crowdsensing,and the design of efficient task assignment mechanism has attracted much attention by many researchers.At present,researches on task assignment for crowdsensing mainly involve in the following requirements,such as time,location,quality and single ability of tasks and users participating in sensing tasks.But there lacks further study on task assignment problems that require diversity abilities.In practice,there are some complex cognitive tasks,which require users with multiple capabilities to collaborate.Focusing on the aforementioned observation,the paper studies the problem of crowdsensing tasks assignment with the constraints of capacities diversity.The main contributions of this paper include the following two aspects.(1)Propose task assignment algorithms with consideration of additional costs.When a user performs multiple sensing tasks,the cost that his/her smart device open the sensor and connect to the wireless network is independent of the number of tasks,while the cost of collecting,saving,and transmitting the sensing data is related to the specific task.For this reason,we divide the cost of completing tasks for users into the basic cost and the additional cost,where the basic cost is independent of the number of tasks completed by users.On the basis of considering various abilities,aiming at maximizing user benefits,we transform the above task assignment problem into the maximum coverage problem,and propose two heuristic algorithms.(2)Propose allocation algorithms for fine-grained tasks.In the case of complex crowdsensing,sensing tasks require not only user collaboration with multiple capabilities,but also a certain number of users for each capability.For this purpose,we introduce the concept of fine-grained tasks,and propose a cost calculation method for users to complete multiple fine-grained tasks.With the goal of minimizing users costs,we transform the problem of fine-grained task assignment into the minimum weighted set coverage problem.Based on this,we design two heuristic task assignment algorithms.Finally,the simulation experiments are carried out for the task assignment algorithm proposed above.The effectiveness and efficiency of the proposed algorithm are verified by comparison with other algorithms,and the proposed algorithms are applicable to the task assignment of various crowdsensing scenarios.The research content of this paper has a great practical significance to the promotion and use of the crowdsensing system.