Research On Collaborative Task Processing Mechanism For Wireless Sensor Networks

Nowadays, many emerging applications for wireless sensor networks necessitate sensing, processing and sending mass data by the sensor nodes but they are associated with real-time requirements. However, the node has limited resources and it could be unable to achieve the goals of the applications if the tasks are carried out by a single sensor node. Multi-node collaborative task processing is an effective solution to the conflict between the application requirements of the networks and the limited resource of the nodes. Using the collaborative strategy, the tasks are executed in parallel on different nodes and the network resources are utilized fully through the task decomposition, task description, task allocation, and task scheduling, by which the application requirements are satisfied and the network lifetime is prolonged.In this paper, multi-objective particle swarm optimization algorithm is used to optimize the task allocation and scheduling to improve the performances of the task allocation and scheduling solutions. At the same time, energy-efficient optimized methods for the task processing of the nodes are also designed. The main contents and contributions of this paper are as follows:1. A new state transition method is proposed by designing one-point crossover and duplication-based mutation operation to transfer the task allocation and scheduling solutions to better states. The new solutions derived from the state transitions satisfy all the requirements and limitations.2. An algorithm based on an improved multi-objective particle swarm optimization algorithm is proposed for the optimized task allocation and scheduling in the applications for wireless sensor networks. Some initial solutions are generated at random and iterative optimization is carried out based on them in the solution space according to the proposed state transition method. Furthermore, weighted entropy based technique for order preference by similarity to ideal solution is used to guide the process of optimization. Simulation results validate that the algorithm implements easily but searches effectively and can obtain multi-objective optimized task allocation and scheduling solutions.3. A low power real-time dynamic voltage scaling algorithm is proposed. The created single task processing model is used to reflect the interferences of associated communication events on the processing of the tasks, and an optimal intra-task voltage scheduling strategy based on control-flow information is introduced to obtain the ideal operational frequency (or voltage) of each basic block in the model. The slack time between the task finish time and deadline is then effectively eliminated by creating frequency splitting method. Performance is evaluated through simulations and simulation results show the effectiveness of the algorithm on reducing the nodes energy consumptions for task processing.4. An improved safe intra-task voltage scheduling algorithm based on scenario detection is proposed. The ranges of few parameters in the program of the task are used to describe scenarios which are detected in the process of the task execution, and some of the remaining execution path can be accurately predicted in the on-line mode which can optimize the voltage scheduling. Furthermore, the most reasonable points in the program to detect scenarios are searched out by the proposed scenario detection point positioning algorithm. Performance is evaluated through simulations and simulation results show the effectiveness of the algorithm on reducing the nodes energy consumptions for task processing.