基于改进的PSO进化神经计算进行苹果颜色快速分级

为克服在苹果颜色分级中存在的速度慢、误差大等缺点,基于再现群智能的粒子群进化算法和神经计算技术,提出了一种新颖、快速的智能分级方法,即首先通过计算机视觉技术获取苹果表面颜色的色度,并提取其特征;然后采用改进的带自适应惯性权值的粒子群优化算法训练神经网络结构,最后用训练好的神经网络进行苹果颜色分级。实际应用表明该方法切实可行且效果显著,不仅分级速度快,而且分级正确率高达98%以上。     

Weather-Centric Rangeland Revegetation Planning

Invasive annual weeds negatively impact ecosystem services and pose a major conservation threat on semiarid rangelands throughout the western United States. Rehabilitation of these rangelands is challenging due to interannual climate and subseasonal weather variability that impacts seed germination, seedling survival and establishment, annual weed dynamics, wildfire frequency, and soil stability. Rehabilitation and restoration outcomes could be improved by adopting a weather-centric approach that uses the full spectrum of available site-specific weather information from historical observations, seasonal climate forecasts, and climate-change projections. Climate data can be used retrospectively to interpret success or failure of past seedings by describing seasonal and longer-term patterns of environmental variability subsequent to planting. A more detailed evaluation of weather impacts on site conditions may yield more flexible adaptive-management strategies for rangeland restoration and rehabilitation, as well as provide estimates of transition probabilities between desirable and undesirable vegetation states. Skillful seasonal climate forecasts could greatly improve the cost efficiency of management treatments by limiting revegetation activities to time periods where forecasts suggest higher probabilities of successful seedling establishment. Climate-change projections are key to the application of current environmental models for development of mitigation and adaptation strategies and for management practices that require a multidecadal planning horizon. Adoption of new weather technology will require collaboration between land managers and revegetation specialists and modifications to the way we currently plan and conduct rangeland rehabilitation and restoration in the Intermountain West.     

双向电液比例张力绞车的直接自适应鲁棒控制

设计了基于比例溢流阀调压回路的双向电液比例张力绞车,并进行了理论分析。针对液压系统的不确定性因素以及卷绕机构参数变化对张力控制的影响,提出了直接自适应鲁棒电液张力控制方法,并用张力负载模拟试验系统进行了研究。理论分析与试验均表明,对于具有不确定性因素的张力跟踪控制,所设计的电液张力绞车及其控制策略,具有良好的控制性能。     

A Method of Adaptive Fuzzy Classfication on Rockmass Blast Ability

This paper presents a series of targets which are usefull for engineering.The paper also presents an adaptive weight function according to the fail mechanism of rockmass and a fuzzy aynthetic evaluation medel on rockmass blast ability.The model is applied to evaluate the blast ability of rockmass in an engineering and the result shows that the model is effective.     

实现农机技术培训的现实途径

提出了与时俱进、因事而立、以人为本的农机技术培训适应性原则，以及认真调查、制定规划、编好教材、选好基地、搞好“四结合”、静态培训与动态培训相结合、“长线”与“短线”相结合的新形势农机技术培训的现实途径。     

Signatures of positive selection for local adaptation of African native cattle populations: A review

Cattle are central to the lives and diverse cultures of African people. It has played a crucial role in providing valuable protein for billions of households and sources of income and employment for producers and other actors in the livestock value chains. The long-term natural selection of African cattle typically signals signatures in the genome, contributes to high genetic differentiations across breeds. This has enabled them to develop unique adaptive traits to cope with inadequate feed supp...     

自适应巡航系统在汽车中的应用

阐述了汽车自适应巡航系统的作用、工作原理和关键技术,介绍了其在国内外的研究应用情况,并对自适应巡航系统未来发展趋势作了探讨.     

静止无功补偿器的智能自适应PID控制器设计

将神经网络和模糊控制与有着广泛应用ＰＩＤ控制相结合,设计了一种静止无功补偿器的智能自适应ＰＩＤ控制器。利用神经网络实现系统模型辨识,采用模糊逻辑和神经网络相结合对ＰＩＤ控制器参数动态寻优。使ＳＶＣ的控制既具有模糊控制的简单,有效的非线性控制作用,又具有神经网络的自学习,自适应能力。     

一种新型的流媒体系统设计与实现

介绍了一种适用于网络带宽不稳定环境下的高效、可扩展、自适应以及鲁棒性高的视频流压缩与传输技术,并以此为基础最终实现了一个流媒体系统。该系统由流媒体压缩和网络传输控制两部分组成,详细介绍了这两部分所采用的关键技术。