As optical lithography advances into the 65nm technology node and beyond, minimum feature size outpaces the lithography wavelength. As a result, mask/wafer manufacturing yield improvement and cost reduction have been widely accepted as key factors for aggressive technology scaling. This thesis is concerned with the following four manufacturability/manufacturing problems. Fracturing: Mask manufacturing for the 90nm and 65nm nodes increasingly deploys variable shaped beam mask writing machines. The pervasive use of OPC leads to dramatic increase in the number of thin trapezoids, which significantly decrease the mask manufacturing yield. This thesis suggests an optimal integer linear programming based fracturing approach and a fast heuristics which substantially reduce sliver count in comparison to leading commercial fracturing tools. MPW: Multiple project wafers (MPW) provide an attractive mask manufacturing cost reduction solution for low-volume production designs by sharing the cost of mask tooling among up to tens of designs. This thesis proposes a comprehensive MPW flow aimed at minimizing the manufacturing cost which includes (1) multi-project reticle floorplanning, and (2) wafer shot-map and dicing plan definition. PSM: In the context of wafer manufacturing, Alternating-Aperture Phase Shift Masking (AAPSM) will be used to image critical features on the polysilicon layer at smaller technology nodes. This technology imposes additional constraints on the layouts beyond traditional design rules. Phase conflicts have to be detected and removed to enable the use of AAPSM. This thesis has two key contributions: (1) a new computationally efficient approach to detect a minimal set of phase conflicts, which when corrected will produce a phase-assignable layout; (2) a novel layout modification scheme for correcting these phase conflicts in standard- cell blocks. Redundant Vias: Finally, a large part of wafer manufacturing yield loss is due to via voids, which can be relieved by redundant vias insertion or via doubling. This thesis proposes perfect matching based post -route via doubling which achieves optimum yield improvement. Redundant interconnects or "short loops& quot; are introduced to maximize the number of doubled vias. Experimental results show that near 100% via doubling coverage can be achieved with simultaneously optimal redundant via and short loop insertion in the post -route stage

Lithium ion batteries (LIB) are regarded as one of the near-term solutions to the gloable energy shortage and have drawn increasing attentions. To be implemented in the large-scale high-power system, performance requirements are raised especially from the aspects of energy/power density, cycling life and safety issues, therefore further LIB material and system developments are necessary. Among Li-intercalation compounds, high voltage spinel compound LiM₀.₅Mn₁.₅O₄ and high capacity layered lithium-excess oxide compounds Li[NixLi₁/₃-₂x/₃Mn₂/₃-x/₃]O₂ are of great interests. On the other hand, the poor rate capability and cycling performances are preventing these materials from commercialization. In this dissertation, systematic studies are performed on these two types of candidate materials using first principles method. The dissertation can be divided into two parts. In the first part, cation charge ordering in LiMn₂O₄ and its effects on Li diffusion in LiMn₂O₄ spinel are investigated. The conclusions are used as the criteria to pre-screen the dopant M in LiM₀.₅Mn₁.₅O₄ by computation. Cu is found to be able to reduce the Li diffusion barriers and a new type of bi- doped spinel LiNixCuyMn₂-x-yO₄ (0

Most Chinese words are compounds formed through the combination of meaningful characters. Yet, due to compositional complexity, it is poorly understood how this combinatorial process affects the access to the whole-word meaning. In the present study, we turned to the recent development in compositional distributional semantics (Marelli et al., 2017), and employed a deep neural network to learn the less-than-systematic relationship between the constituent characters and the compound words. Based on the compositional representations derived from the computational model, we quantified compositionality as the degree of overlap between the compositional and the lexicalized representations as well as the degree of distinctness of the compositional representation. We observed that these two compositional attributes can affect compound recognition over and above the effects of constituent character features and compound features. Moreover, we found that this effect was increasingly stronger when holistic access to the compound meaning became more challenging. These findings therefore, from a computational perspective, provided new evidence for the combinatorial process involved in Chinese word recognition, which also shed light on the universal process of compound comprehension.

Children face the problem of extending a limited spatial lexicon to potentially infinite spatial situations. Previous work has examined how spatial semantic categories may be formed in child development, but it is unclear how children extend these categories to novel situations over the developmental time course. Drawing on cognitive linguistic theories of category extension, we present a framework that models the incremental extension of spatial relational words to novel situations through time. We describe a longitudinal dataset and computational analyses for investigating the extension of spatial word meanings in a developmental setting. Our preliminary results suggest that the formation of spatial categories takes place through an exemplar-based process of chaining, similar to the process underlying the growth of linguistic categories in history. Our work offers opportunities to explore the connection between ontogeny and phylogeny in the process of word meaning extension.

In this paper, we propose a cost-effective localization solution for land vehicles, which can simultaneously adapt to the uncertain noise of inertial sensors and bridge Global Positioning System (GPS) outages. First, three Unscented Kalman filters (UKFs) with different noise covariances are introduced into the framework of Interacting Multiple Model (IMM) algorithm to form the proposed IMM-based UKF, termed as IMM-UKF. The IMM algorithm can provide a soft switching among the three UKFs and therefore adapt to different noise characteristics. Further, two IMM-UKFs are executed in parallel when GPS is available. One fuses the information of low-cost GPS, in-vehicle sensors, and micro electromechanical system (MEMS)-based reduced inertial sensor systems (RISS), while the other fuses only in-vehicle sensors and MEMS-RISS. The differences between the state vectors of the two IMM-UKFs are considered as training data of a Grey Neural Network (GNN) module, which is known for its high prediction accuracy with a limited amount of samples. The GNN module can predict and compensate position errors when GPS signals are blocked. To verify the feasibility and effectiveness of the proposed solution, road-test experiments with various driving scenarios were performed. The experimental results indicate that the proposed solution outperforms all the compared methods.

We study signal integrity effects on statistical timing analysis, e.g., interconnect and gate delay variations induced by crosstalk aggressor alignment, i.e., difference in signal arrival times in coupled neighboring interconnects. Such effects bring significant source of variation, and must be taken into account in statistical timing analysis. We establish a functional relationship between signal propagation delay and crosstalk alignment by deterministic circuit simulation, and derive closed form formulas for statistical distributions of output signal arrival times. Our proposed method can be smoothly integrated into a static timing analyzer, which runtime is dominated by sampling deterministic delay calculation, while probabilistic computation and updating take constant time. Our experimental results on 70nm technology global interconnect structures and 130nm technology industry designs show that lack of statistical crosstalk alignment consideration could lead to up to 114.65% (71.26%) differences in interconnect delay means (standard deviations), and 159.4% (147.4%) differences in gate delay means (standard deviations), while our method gives within 1.28% (3.38%) mismatch in interconnect output signal arrival time means (standard deviations), and within 2.57% (3.86%) mismatch in gate output signal arrival time means (standard deviations), respectively.

Pre-2018 CSE ID: CS2007-0883