Aims:With the rapid expansion of road networks,the conflict between transportation development and ecological conservation has become increasingly prominent,garnering significant attention from ecologists.Research on road ecology has indicated that roads can fragment animal habitats,reduce connectivity between populations,and severely impact biodiversity.Among these impacts,animal vehicle collisions(i.e.,roadkill)represent the most direct and significant threat to wildlife. Methods:This research investigated the incidence of animal vehicle collisions in Nanjing,a mega city in East China.From November 2020 to October 2021,transect survey method was employed to investigate roadkill incidents across roads of different administrative levels in Nanjing.Regression analysis was used to explore the probability of roadkill incidents,and models were established to further clarify the impact of roads and surrounding environmental factors on the probability of animal roadkill. Results:Our findings revealed that:(1)A total of 293 roadkill incidents involving 21 species were recorded during the survey period.Mammals accounted for 46.42％of the incidents,with cats(Felis catus)and dogs(Cams lupus familiaris)being the most common,while birds accounted for 48.81％,with the Chinese Blackbird(Turdus mandarinus)and Eurasian Tree Sparrow(Passer montanus)being predominant.(2)The proportion of arable land(dryland and irrigated farmland)and buildings(buildings and structures)in natural landscape significantly influenced the probability of roadkill incidents.Specifically,within a 250 m radius around the road,the coverage proportion of arable land and buildings was significantly negatively correlated with the occurrence probability of roadkill incidents.As cultivated land coverage increased,the probability of roadkill of birds decreased,while the probability of mammals'roadkill increased,showed completely opposite trends.(3)The road characteristic affected the probability of animal vehicle collisions;roads with isolation belts were associated with an increased probability of roadkill incidents.(4)Linear features(distance to the nearest river,the nearest viaduct and the nearest main road)also affected the incidence of bird's roadkill events.Birds had fewer vehicle collisions the farther they were from rivers,but more collisions while the farther they were from viaducts and main roads. Conclusion:This study represents the first periodic investigation into terrestrial vertebrate roadkill in Nanjing.It preliminarily identifies the natural landscape and road factors that influence animal roadkill in Nanjing,thereby providing foundational knowledge to inform strategic decisions regarding regional road network optimization,infrastructure enhancements,ecological corridor establishment,and management strategies.

Group-size effects, whereby antipredator vigilance decreases as group size increases, are widely reported in mammals and birds but a meta-analysis has only been conducted in birds. We systematically reviewed the literature on mammalian group-size effects, estimated the effect sizes in each study, and conducted a phylogenetic meta-analysis. We obtained 296 effect sizes from 97 species belonging to 10 Orders and 26 Families. Overall, effect sizes indicated a moderate negative effect of group size (r = −0.44), but 43% of the effect sizes were compatible with a null effect of group size. There was significant heterogeneity in effect sizes. Weaker effect sizes occurred when vigilance was measured as a frequency or a duration rather than as a percentage of time spent vigilant, when measured in closed habitats, during the reproductive season, and in mixed-sex groups or during times when juveniles were absent. We infer a “file drawer problem” because there were relatively few studies with smaller sample sizes reporting small group-size effects. The results confirm the importance of group size in explaining variation in mammalian vigilance but also suggest which a substantial amount of variation remains unexplained. We suggest that future studies should aim to study mammalian group-size effects by quantifying the percentage of time allocated to vigilance rather than lower-power methods such as frequency or duration of vigilance.

In this study, the improved Aesop’s fable paradigm—a series of experiments originally used to test whether some animals understand the causality associated with water replacement—was used to explore the cognitive ability of Azure-winged magpies ( Cyanopica cyanus ). Experimental results on causal cue tasks showed that the Azure-winged magpies prefer water-filled tubes over sand-filled tubes, heavy objects over light objects, and solid objects over hollow objects. However, they failed to notice the diameter and water level of the tubes. They also failed to pass the counterintuitive U-shaped tube task in arbitrary cue tasks. Our results demonstrated that Azure-winged magpies have a certain cognitive ability but not an understanding of causality, a characteristic comparable to that of other corvids. Moreover, Azure-winged magpies exhibited the ability of training transfer and analogical problem solving from the perspective of cognitive psychology. We believe that object-bias has little effect on Azure-winged magpies in this study. We can conclude that the Azure-winged magpies partially completed the tasks by trial-and-error learning.

Visual tracking plays an important role in computer vision research in these years. Multi-kernel correlation filter has demonstrated its outstanding advantage via introducing high level representation from multi-kernel. However, the unskillful selection of multi-kernel inevitably brings redundancy and noise within learning and updating procedure, which significantly affects the accuracy of tracking. A large margin multi-kernel tensor correlation filter for visual tracking (LMKCF) is proposed in this paper. The LMKCF mainly mitigates the redundancy and noise of multi-kernel correlation filter in learning and updating from two aspects with the low rank tensor learning to establishes a prospective learning and updating strategy. And the optimization problem can be solved effectively by the alternating direction method of multipliers (ADMM) method. Last, we validate the proposed tracker with the multi-kernel representations based on OTB benchmark to demonstrate the superiority of the method.

The energy-saving technology for hybrid optical wireless broadband access network (WOBAN) system is an urgent issue nowadays. In the optical wireless broadband access network, the energy consumption of the wireless network occupies the vast majority of the total energy consumption of the network. In traditional energy saving scheme, less consideration is given to the energy saving of the wireless network side. This thesis proposes an optical and wireless hybrid energy saving scheme based on load transfer (LTES) for a typical commercial and residential hybrid building scenario. The scheme puts the low-load network nodes into a sleep mode, and transfer the part of the load to other active nodes through re-routing. Mathematical derivation and experiment show that the scheme can effectively improve the average energy saving efficiency compared with the traditional scheme. When the household users account for 25% and 50% of the total users, respectively, the average energy efficiency is increased by about 11.3% and 20.4% respectively during the day when the total load of the users is large.

A blocking-triggered defragmentation mechanism based on the selected path (SP-D) is proposed to solve spectrum fragments in elastic optical networks (EONs). The proposed mechanism analyzes the arriving traffic requirement as well as the current spectrum condition, and then takes the effectiveness and the complexity of spectrum defragmentation into consideration simultaneously. When service blocking is detected, a path from source node to destination node is selected according to the proposed policy and then the spectrum reallocation is performed sequentially based on the selected path to ease the blocking situation. Theoretical analysis and numerical simulation results show that compared with the previous mechanisms, the proposed SP-D mechanism can reduce the network latency as 7.67% and decrease the bandwidth blocking probability (BBP) as 5.3%, respectively.

Routing and spectrum allocation (RSA) is one of the key problems in EONs since bandwidth fragmentation generated by establishing the connection dynamically. In this work, we propose a routing and spectrum allocation algorithm based on Maximum Spectrum Completeness (MSC) which selected the completely contiguous frequency of the light-path and avoiding generating single fragments in the route. The probable processes are to first choose the appropriate routes for the request based on $k$ shortest path algorithms. Secondly, selecting the spectrum with the smallest remaining available frequency slot in the Exact Fit scheme among alternative paths respectively. Finally, spectrum allocation is performed according to the spectrum completeness formula to calculate the scheme with the largest continuously available spectrum blocks. Simulation results prove that MSC algorithm provides better performance than traditional scheme, and the BBP is reduced approximately 19.8% than FF scheme, furthermore, over 7.1% than EF scheme.

Spectrum fragment is one of the most important issues in elastic optical networks (EONs). To solve the shortcomings of traditional defragmentation mechanisms, a policy-based comparison-triggered defragmentation mechanism (P-CTD) is proposed. The improved mechanism will analyze the arrived traffic request and current spectrum resource situation firstly, then compare and judge whether to trigger spectrum defragmentation according to preset policy. Theoretical analysis and numerical simulation results show that compared with the traditional mechanisms, the proposed P-CTD mechanism can reduce the bandwidth blocking probability (BBP) as 8.73% and improve the resource utilization ratio as 15.26%, respectively.

A Multi-Domain Awareness Routing algorithm is proposed to reduce the energy consumption in elastic optical networks. By adopting the fragmentation metrics both of the spectrum domain and spatial domain, the blocking probability of elastic optical networks can be reduced. When new services arrive, they are prioritized to work on existing links to reduce new optical paths and working network components. Then the holding time of existing services is fully considered in order to select the candidate route with a shorter working time and thus improve energy consumption. The simulation results show that the proposed algorithm can effectively reduce the network energy consumption and hence, result in overall lower connection blocking probability.

弹性光网络(EON)中动态光路的建立与拆除容易产生频谱碎片,不合理的频谱分配策略会导致更高的带宽阻塞率.针对此问题提出了一种基于频谱候选窗(SCW)的EON改进频谱分配方案,对到达的业务寻找可用的SCW,按空闲频谱块连续度进行筛选以后再进行频谱分配.理论分析和数值仿真结果表明:相比于EF(Exact Fit)频谱分配策略,改进方案的带宽阻塞率性能最优可降低8.76％.

为了降低日益增长的接入网能耗,研究了无源光网络(PON)系统中的节能技术,在传统节能策略的基础上,提出了一种基于OLT缓存的节能策略.该方法通过在OLT侧设置一个缓存器,缓存来自核心网的下行数据业务,根据下行数据业务的大小是否超过预设的门限值来判定是否向相应的ONU发送下行数据.理论分析和仿真实验表明:相对传统无节能方案,该方案可以节能75％,且在下行链路负载较低时,下行分组平均时延有明显改善.

弹性光网络(EON)中业务频谱分配需要同时满足子载波频谱连续性约束和链路一致性约束,当剩余可用频隙不满足约束条件时,易导致请求业务阻塞.提出了一种基于不相交连接组优化的频谱分配方案,对到达的业务按照带宽需求和路由长度分配不相交连接组和相交连接组,分别采用对应分配策略以获得更多连续可用频隙,避免产生频谱碎片.理论分析和仿真结果显示:在100～600Er1内,相比传统分配策略,改进方案获得的连续可用频隙比改善了约8.6％～12.7％;在网络负载为600Er1时,相比First Fit和First-Last Fit方案,改进方案的带宽阻塞率分别改善约13.8％和6.8％.