| In recent years, along with the increases of heterogeneous, dynamic and complex, the resource management has become the main factor that restricts the service efficiency of intelligent space so that the research of resource management in intelligent space is significant and valuable. In this thesis, firstly, different resource management systems are designed according to the specific characteristics of resources in intelligent space. Secondly, to solve the problems of resource conflict in intelligent space, the corresponding resource scheduling strategies are proposed. Lastly, to solve the problem of service selection in platform based on intelligent space technology and cloud computing technology, a service selection approach is proposed. The main works of this thesis are listed as followed.The framework of intelligent space resource management system is designed. In this thesis the resources are divided into two categories:physical resources and virtual resources. Physical resources are divided into electrical equipment resources and robot resources. RM04 module and STM32 MCU (Microcontroller Unit) are used to construct WIFI control network of the intelligent space. The WIFI control network can achieve the remote management and control of equipment resources. The UPnP (Universal Plug and Play) technology is used to establish a ROS (Robot Operating System) robot middleware for robot resources. The ROS robot middleware realizes the loose coupling and zero configuration between the intelligent space and ROS robot. It also can improve expansibility and compatibility of system. Based on UPnP and WCF (Windows Communication Foundation) technology, an intelligent space resource management system oriented to service is designed in this thesis. Virtual resources in intelligent space can be managed and integrated through this system. An unified human-computer interaction Web interface is designed, which helps users to manage and monitor the resources of the whole intelligent space in real time.The resource conflict problems in intelligent space are explored and the resource scheduling strategies are proposed. A method for dynamic matching of equipment based on fuzzy centralized sorting is proposed to solve equipment resource conflict problem. High priority tasks can be implemented as soon as possible through this method, meanwhile the idle entity resources are effectively utilized. In this thesis, a ROS robot service node management system is designed for node management problem of ROS service robots in a distributed environment. This system can automatically start and close the ROS node in the distributed environment based on the information of ROS devices. The system reduces the burden of the robot by management of the ROS services. A scheduling strategy based on greedy algorithm and genetic algorithm is proposed for resource conflict and load balancing problems of virtual resources. The resource scheduling strategy is mainly divided into two kinds of situations:when the virtual resources are inadequate, the virtual resource conflict problem is abstracted into the 0-1 multi knapsack problem, and the genetic algorithm is used in this case. On the other hand, when the virtual resource of intelligent space is sufficient, the greedy algorithm is used for virtual resources allocation. This virtual resource scheduling strategy can effectively improve the utilization of virtual resources and reduce the service execution time.Finally, the framework of the combination of intelligent space platform and cloud computing platform is studied and designed in this thesis. A service selection approach based on FAHP (Fuzzy Analytical Hierarchy Process) and SVD (Singular Value Decomposition) is proposed for service selection in this mixing platform. In the case of insufficient service information, this approach is able to select appropriate services adaptively from similar services in intelligent space and cloud computing platform by integrating environment information and service characteristic. |
PDF Full Text Request