摘要
针对永磁球形步进电机的特殊结构和旋转特点,首先进行位置检测系统的数学建模。对于转子球体目标节点的高密度计算问题,采用多任务并行混沌搜索的方法,同时采用适合并行计算的片上网络(NOC)硬件结构来实现,讨论当前任务下功耗优先的NOC映射问题。仿真结果表明多任务并行搜索的效率要比串行搜索效率高得多,基本达到预期目的,为今后的FPGA设计和位置检测的实时性研究提供依据。
On the basis of the special structure and rotational character of the permanent magnetic step motor,set up the mathematically model the system of detecting the rotor orientation.To calculate the goal node of the spherical rotor for a great scale,multi-task parallel chaos search was used,at the same time the hardware structure of network on chip（NOC）fitting parallel calculaton was used,and discussed the current task NOC mapping of power-consumption-first.Simulation results show the efficiency of multi-task parallel search is much higher than the efficiency of serial search,which meet the expectant demand and lay foundation for the furture FPGA design and the real-time research of detecting the rotor orientation.
出处
《微电机》
北大核心
2010年第6期43-47,共5页
Micromotors
基金
国家自然科学基金项目(50377010)
安徽省自然科学基金项目(03044103)
关键词
球形电机
NOC
映射
并行搜索
效率
Spherical motor
NOC
Mapping
Parallel Search
Efficiency
