摘要
为了利用最小资源解决汽车制动信号的大数据量处理和识别问题,提出了一种基于8位微控制器的汽车制动信号快速识别算法。在保证准确识别汽车制动信号的基础上,采取一种新颖的数据逻辑存储结构,将运算量降低到最小,一次采样只需进行1次加法运算,128次采样只需1次减法运算,即可实现信号平滑和识别,并将数据存储空间降低到233字节,不必扩展存储器,大大地降低了系统成本和功耗。该算法基本上不受平滑滤波等参数的影响,具有很好的鲁棒性,已成功运行于MCS51和AVR微控制器组构的处理系统上。
To most efficiently processing automobile massive braking signals, a fast algorithm based on 8 bit MCU architecture is proposed for detecting automobile braking signals. In this algorithm, a novel logic storage structure is implemented to reduce the computation complexity of signal processing to minimum, without losing the preciseness of automobile brake signal detection. In the proposed algorithm, for detecting and smoothing signals, only one addi- tion operation is needed for one sampling operation and one subtraction operation for 128 sampling operations. Ad- ditionally the storage space has been reduced to 233 bytes without the need for extended memories. It can signifi- cantly reduce the system cost and the power consumption. The proposed algorithm is robust due to its insensitivity to smoothing filter parameters and has been successfully implemented on the circuit processing systems based on MCS51 and AVR MCU architecture.
出处
《科学技术与工程》
2008年第4期908-913,共6页
Science Technology and Engineering
基金
国家自然科学基金项目(60472006)
