Research On Centralized Scheduling In Industrial Wireless Sensor Networks

Wireless Sensor Networks (WSNs) have become an important Industrial Measurement & Control (IMC) technology, with some characteristics including low cost, ease of deployment, intelligence, etc. In order to improve network controllability & manageability and meet the requirements of industrial applications, Industrial Wireless Sensor Networks (IWSNs) are designed to be centralized control-the manager is responsible for network communication and resource scheduling.This dissertation studies a scheduling pattern based on the separation of control and data planes for centralized IWSNs. Based on this, it focuses on scheduling algorithms on the network layer and the data link layer. The main contributions of this dissertation are as follows:Centralized scheduling protocol in IWSNs is proposed. This paper studies logic control function of the control planes, forwarding process of the data plane, cross-layer scheduling of protocol stack and network resource abstraction & scheduling strategies. Based on TDMA, control channels and data channels are separated by clear channel assessment mechanism. The study further designs centralized cross-layer scheduling protocol. Experiment results show great advantages compared with traditional WSNs in transmission delay and packet success rate.Resource-aware graph routing algorithms are proposed. Industrial standards adopt graph routing at the network layer. Graph routing algorithms are generated by resource-aware routing metric algorithms for broadcast, uplink and downlink communication respectively. The routing metric algorithms depend on multiple metrics and different fault-tolerant mechanisms. In consideration of routing metrics and communication requirements, paths are selected by manager and routing graph can be constructed. Simulations evaluate the packet reception rate and the network overhead of resource-aware graph routing algorithms and analyze the affect of resource-aware graph routing algorithms by link loss rate.Timeslot reuse based multipath transmission scheduling algorithms are proposed. In multipath transmission, scheduled network resources increase along with the increment of paths. In order to improve resource utilization, we present an optimized scheduling method by timeslot reuse, which optimizes the path selection and make conflict-free nodes share communication resources. Timeslot reuse based multipath transmission scheduling algorithms can be achieved by path tree establishment, traversal and optimization. The results of simulations show that the number of scheduled timeslots is reduced without decreasing the packet reception rate.Multi-constrained based latency-optimal scheduling algorithms are proposed. Industrial standards support frequency hopping and the manager should schedule both timeslots and channels. This dissertation formulates resource scheduling problem and theoretically analyzes the minimum number of timeslots with unlimited frequency and limited frequency. With constraint conditions of transmission path and conflict-free scheduling, in above cases packet based latency-optimal scheduling algorithms for timeslots and channels are proposed. Latency-optimal scheduling algorithms are more flexible and scalable, compared with node based heuristic scheduling algorithms. Simulations evaluate the performance of these two algorithms based on the minimum number of timeslots.