Research On TDMA Link Scheduling For Data Collection Sensor Networks

Data collection sensor networks as a new means of data gathering, widely used in all kinds scenarios which need to collect sensing data periodically, such as environmental monitoring, conservation protection, health monitoring. Because of the multi-hop data transfer mode in data collection sensor networks and the limitation of communication range for the nodes, there are plenty of opportunities to data concurrent transmission. TDMA link scheduling greatly improve the network performance by scheduling the nodes’activities with the interference-aware manner.However, due to the sensor node’s hardware resource constraints, limited network bandwidth and node energy, wireless interference, and so on, how to implement a TDMA link scheduling so that it can optimize network performance is still facing many challenges and problems. For these reasons, this thesis carried out the research on TDMA link scheduling from single-channel and multi-channel scheduling. For single channel, this thesis proposed centralized scheduling algorithm based on the network conflict graph and distributed scheduling algorithm based on the cooperation of nodes. In multichannel, this thesis proposed multi-channel link scheduling approach with optimized network energy consumption.For single-channel centralized link scheduling, first of all, taking into account the impact of wireless interference on constructing network conflict graph, this thesis proposed an accurate point-to-point interference measurement approach (InterM, Point-to-Point Interference Measurement). By assigning the conflict-free time slot for interference measurement to nodes, InterM ensures that the nodes are able to measure the radio interference accurately. Secondly, a network conflict graph-based workload-aware link scheduling algorithm (WAGS, Workload-Aware Greedy Scheduling) was proposed to deal with the nodes’uneven initial workload and the energy cost for the nodes" wake-up in the data collection sensor networks. By using workload-aware time slot assignment and a greedy time slots reuse mechanism, WAGS reduces delay and energy consumption. Finally, on the basis of workload-aware, a link scheduling algorithm combined with the data fusion technique (PriorS) was proposed and further reduce the node’s energy consumption. For single channel distributed link scheduling, because centralized scheduling algorithms needs to collect the entire network topology information and construct network conflict graph, and so these algorithms cannot be used in a large-scale data collection sensor networks. This thesis proposed a distributed link scheduling algorithm (Coop. Cooperation-based Scheduling with Consecutive Timeslots Assignment) which combined with a depth-first traversal method, the time slot consultation mechanism, as well as limited continuous time slot assignment. The Coop improves the reliability of data concurrency transmission and energy efficiency. In Coop, however, only one node be able to perform the time slot assignment in every time slot assignment process. This thesis proposed a distributed algorithm with concurrent time slot assignment (CDTS. Cooperation-based Distributed TDMA Scheduling). A round-based time slot assignment and distributed time slot vote mechanism were developed to guarantee the reliability of data collection and energy efficiency and reduce the cost of time slot assignment.For multi-channel link scheduling, since many-to-one communication pattern, consecutive time slots assignment in single-channel cannot minimize the network energy consumption in a low-rate, low traffic load data collection network, the scheduling with unlimited channels (SUC) and scheduling with limited channels (SLC) were proposed to minimize the network energy consumption. The algorithm uses a receiver-based consecutive time slots assignment approach and greedy multichannel allocation technology to optimize network energy consumption with the fewest channels.Future research work includes:developing a centralized scheduling approach combined with the order of nodes’ schedules, delay and energy consumption into consideration; developing distributed time slot adjustment mechanism to adapt to network topology changes quickly with low overhead; optimizing multi-channel scheduling approach with inter-channel and intra-channel time slot reuse to reduce the data collection time.