| dc.description.abstract | Existing smart sensor system often lacks effective management mechanisms for multi-sensor system, resulting in a large burden on the processor, low efficiency and high power consumption; while it’s difficult for smart sensor based on processor-based fieldbus protocol and IEEE1451 to achieve the goal of real-time, reliability, low-cost and so on. This paper presents an innovative intelligent sensor network architecture that can effectively reduce power consumption, improve real-time performance, and has a better management performance. This system consists of smart sensor nodes, smart sensor network gateway, both hardware realized , and a client monitoring program base on C++ builder.
The system we designed meets the definition of smart sensor, such as non-linear calibration, self-compensating, self-inspection and self-diagnosis. What’s more, this system further meets the definition of smart sensor network, like nodes added detection, nodes removal detection, signal acquisition, fault detection, mode control, ID configuration. System uses Modbus hardware modules as the network communication infrastructure. In addition, we adopted the CRC, ARQ and Hamming code error correction mechanism to improve the reliability of real-time communication. In this study, we used Hierarchical Robotic Discrete-Event Modeling for the hardware implementation of all embedded functional modules. Additionally, the baud rate, Modbus functions and hardware modules can be easily adjusted and reconfigured, raising system performance and flexibility. Finally, we implemented FPGA prototype system by using Altera Cyclone III-C25 FPGA and DuPont lines. Under this short-range communication condition, the maximum system baud rate speeds up to 7.06Mbps, the node response time is only 73.8us, and data transfer success rate achieves 100%, with no data losing phenomenon. The system is in line with real time and reliability of system requirements. | en_US |