Real-Time Performance Reseach On Embedded Meteorologic Acquisition System And Network

With the accelerating meteorological service development and the improving accurate and timely forecast alert service, Nation offers to design and popularize the second-generation meteorologic acquisition stations for an important social significance and a broad market prospect. The second generation station bases on embedded system technology and CAN bus technology, designs according to the "the main collector + external network + sub-collector + sensor + peripherals" structure, aims at high precision, high stability, easy maintenance, low power consumption, easy extension and real-time remote monitoring.This paper researches on the automatic weather station of ARM processor, embedded Linux operating system and CAN network as the research object, builds cross-compiling development environment, transplants root filesystem and kernel, designs CAN driver program and data transceiver program. This paper chooses a suitable meteorological collection system network CAN network, proposes an improved mix scheduling algorithm (MTS) reforming to improve the real-time performance, applied the algorithm in the system, made overall system performance improved. This paper analyses the real-time performance of Linux 2.6.24 kernel the system adopts in, optimizes this system clock granulator in-depth, and applys real time algorithm LSF to improve system performance. It uses open source system test tools LMbench to test and compare the context switch time before and after transformed. Experiments have proved that the context switch time is obviously less than before.