Search

Curb detection is an essential function of autonomous vehicles in urban areas. However, curbs are difficult to detect in complex urban environments in which many dynamic objects exist. Additionally, curbs appear in a variety of shapes and sizes. Previous studies have been based on the traditional pipeline, which consists of the extraction and aggregation of hand-crafted features that are then fed to classifiers. However, this sequential process is inefficient and designing the hand-crafted features is a complex process. Recently, this kind of process has been replaced by Deep Neural Networks (DNN), in which classifiers and features are learned from large-scale data. Very few works have exploited DNN for the curb detection problem. Most works use multi-modal sensor-based methods that combine images and accumulated 3D point clouds from LIDAR. However, these approaches require synchronization and calibration between sensors. In addition, they do not quantify the uncertainty of their predictions for autonomous system safety. In this paper, we present a two-stage DNN-based curb detection method that includes uncertainty quantification. An autoencoder-based network predicts the curbs, and then conditional neural processes rectify the predictions with uncertainty estimations. The experimental results show that our approach achieves high accuracy and recall in complex areas. We also constructed a large-scale dataset to create benchmarks consisting of approximately 5,224 scans with bird’s-eye view labels collected from urban areas. To the best of our knowledge, there are no public datasets for DNN-based curb detectors. The benchmarks and datasets are publicly available at https://github.com/YounghwaJung/curb_detection_DNN.

In recent years, artificial intelligence (AI) technology has proliferated rapidly and widely into application areas such as speech recognition, health care, and autonomous driving. To increase the capabilities of AI more powerful systems are needed to process a larger amount of data. This requirement has made domain-specific accelerators, such as GPUs and TPUs, popular; as they can provide orders of magnitude higher performance than state-of-the-art CPUs. However, these accelerators can only operate at their peak performance when they get the necessary data from memory as quickly as it is processed: requiring off-chip memory with a high bandwidth and a large capacity [1]. HBM has thus far met the bandwidth and capacity requirement [2] –[6], but recent AI technologies such as recurrent neural networks require an even higher bandwidth than HBM [7]–[8]. While a further increase in off-chip bandwidth can be accomplished by various techniques, it is often limited by power constraints at the chip or system level [9]. Hence, it is essential to decrease demand for off-chip bandwidth with unconventional architectures: such as processing-in-memory. In this paper, we present function-In-memory DRAM (FIMDRAM) that integrates a 16-wide single-instruction multiple-data engine within the memory banks and that exploits bank-level parallelism to provide $4 \times $ higher processing bandwidth than an off-chip memory solution. Second, we show techniques that do not require any modification to conventional memory controllers and their command protocols, which make FIMDRAM more practical for quick industry adoption. Finally, we conclude this paper with circuit- and system-level evaluations of our fabricated FIMDRAM.

Can a robot be a personal dating coach? Even with the increasing amount of conversational data on the internet, the implementation of conversational robots remains a challenge. In particular, a detailed and professional counseling log is expensive and not publicly accessible. In this paper, we develop a robot dating coaching system leveraging corpus from online communities. We examine people's perceptions of the dating coaching robot with a dialogue module. 97 participants joined to have a conversation with the robot, and 30 of them evaluated the robot. The results indicate that participants thought the robot could become a dating coach while considering the robot is entertaining rather than helpful.

Retrosynthesis, of which the goal is to find a set of reactants for synthesizing a target product, is an emerging research area of deep learning. While the existing approaches have shown promising results, they currently lack the ability to consider availability (e.g., stability or purchasability) of the reactants or generalize to unseen reaction templates (i.e., chemical reaction rules). In this paper, we propose a new approach that mitigates the issues by reformulating retrosynthesis into a selection problem of reactants from a candidate set of commercially available molecules. To this end, we design an efficient reactant selection framework, named RetCL (retrosynthesis via contrastive learning), for enumerating all of the candidate molecules based on selection scores computed by graph neural networks. For learning the score functions, we also propose a novel contrastive training scheme with hard negative mining. Extensive experiments demonstrate the benefits of the proposed selection-based approach. For example, when all 671k reactants in the USPTO {database} are given as candidates, our RetCL achieves top-1 exact match accuracy of $71.3\%$ for the USPTO-50k benchmark, while a recent transformer-based approach achieves $59.6\%$. We also demonstrate that RetCL generalizes well to unseen templates in various settings in contrast to template-based approaches., Comment: Accepted to IJCAI 2021. Short version was accepted to Machine Learning for Molecules Workshop at NeurIPS 2020

In this paper, we present a driverless taxi system for autonomous campus mobility services. College campuses have unique mobility requirements in terms of layout, population, and demand and patterns. It is typically recommended to minimize the presence of private automobiles on campuses due to teaching and research disturbances, visual degradation from parking provision, environmental pollution, and negative health effects. As an alternative to private automobiles, shared mobility systems have been considered for both campus and urban transportation. Conventional shuttle systems suffer from the first and last mile problem. A bicycle and pedestrian friendly policy is not a generalizable solution for all geographic locations and campus layouts. We suggest a driverless taxi service as an alternative point-to-point shared mobility system for campuses. We have demonstrated the feasibility of this service on a 4.5-km campus road at Seoul National University. The service has covered over 10 000 km autonomously since the first public demonstration was made in November 2015.

In LIght Detection And Ranging (LIDAR)-based object detection, accurate object segmentation is of great importance, since segmentation is an essential preprocessing step for other perception tasks, such as classification and tracking. For segmenting objects, most of the previous methods have tried to eliminate the ground first, which typically incurs considerable overhead in computation and inaccuracy in object detection with point clouds gathered by using 3-D LIDARs. However, in many real-time applications, such as automated driving, segmentation should be performed within a specified time, because even a small delay in computation could result in vehicle collisions. In this paper, we propose a real-time and accurate object segmentation algorithm for 3-D point clouds, which does not carry out ground extraction as a first step. In the proposed algorithm, we generate candidate points of objects and find their borders based on the integrated structure of a 2-D grid and an undirected graph, which enables fast processing and yields an accurate segmentation result independent of ground extraction error. In order to enhance segmentation accuracy, we employ Gaussian process, which reduces over-segmentation that separates an object into multiple portions. We apply two types of Gaussian process models to alternately provide cues for merging adjacent over-segmented objects. Experimental results demonstrate that this paper achieves a real-time processing speed and higher segmentation accuracy than previous works in most evaluation metrics. With the application to tracking, we show that the enhanced segmentation accuracy increases the tracking accuracy by 11.4% even in the worst case.

The Li2MnO3 cathode material is vulnerable to complex degradation behaviors during the operation of battery although it has attracted much attention recently due to its potentially large capacity. In this study, we comprehensively examined the degradation process in Li2MnO3, using theoretical density functional computations as well as experimental techniques (in situ X-ray absorption near edge structure spectroscopy, X-ray diffraction, and Raman spectroscopy). Our study reveals that during the delithiation process, the Li ions mixed in the Mn layer are removed together with those in the Li layer, thereby inducing the release of oxygen atoms. The oxygen loss reaction is energetically favorable at the highly delithiated states, and it can reduce the plateau voltage in the charging curve. Such oxygen loss was observed during or even before the second cycle and furthermore it accelerates the phase transformation of the layered structure to a spinel one. Our results also suggest that oxygen release can be prev...

Ni-rich layered oxides are considered to be a promising cathode material with high capacity, and their surface structure should be extensively explored to understand the complex associated phenomena. We investigated the surface stability and morphology of LiNiO2 as a representative of these materials by using density functional theory calculations. The results reveal that the Li-exposed surfaces have lower energies than the oxygen surfaces, irrespective of the facets, and the Ni-exposed ones are the least stable. The equilibrium morphology can vary from truncated trigonal bipyramid to truncated egg shape, according to the chemical potential, whose range is confined by the phase diagram. Moreover, the electrochemical window of stable facets is found to strongly depend on the surface elements rather than the facet directions. Contrary to the stable Li surfaces, oxygen exposure on the surface considerably lowers the Fermi level to the level of electrolyte, thereby accelerating oxidative decomposition of the ...

The development of intelligent vehicle systems has resulted in an increased need for a high-precision road map. However, conventional road maps that are used for vehicle navigation systems or geographical information systems (GISs) are insufficient to satisfy new requirements of intelligent vehicle systems such as autonomous driving. There are three primary road-map requirements for intelligent vehicle systems: centimeter-level accuracy, storage efficiency, and usability. However, no existing researches have met these three requirements simultaneously. In this paper, we propose a precise and efficient lane-level road-map generation system that conforms to the requirements all together. The proposed map-building process consists of three steps: 1) data acquisition, 2) data processing, and 3) road modeling. The road data acquisition and processing system captures accurate 3-D road geometry data by acquiring data with a mobile 3-D laser scanner. The road geometry data are then refined to extract meta information, and in the road modeling system, the refined data are represented as sets of piecewise polynomials to ensure storage efficiency and usability of the map. The proposed mapping system has been extensively tested and evaluated on a real urban road and highway. The experimental results show that the proposed mapping system outperforms conventional systems in terms of the road-map requirements.

Ni-rich layered oxides are attractive materials owing to their potentially high capacity for cathode applications. However, when used as cathodes in Li-ion batteries, they contain a large amount of Li residues, which degrade the electrochemical properties because they are the source of gas generation inside the battery. Here, we propose a computational approach to designing optimal coating materials that prevent gas evolution by removing residual Li from the surface of the battery cathode. To discover promising coating materials, the reactions of 16 metal phosphates (MPs) and 45 metal oxides (MOs) with the Li residues, LiOH, and Li2CO3 are examined within a thermodynamic framework. A materials database is constructed according to density functional theory using a hybrid functional, and the reaction products are obtained according to the phases in thermodynamic equilibrium in the phase diagram. In addition, the gravimetric efficiency is calculated to identify coating materials that can eliminate Li residue...

Precise and robust distance measurement is one of the most important requirements for driving assistance systems and automated driving systems. In this study, the authors propose a new method for providing accurate distance measurements through frequency-domain analysis based on a stereo camera by exploiting key information obtained from the analysis of captured images. Moreover, the proposed method was extensively tested and evaluated on a real urban road, highway and tunnel. Based on these results, the authors show that the proposed method provides more precise distance information in real time compared with conventional algorithms. By applying the authors' methodology to measure the distances of various objects, it can be demonstrated that their algorithm offers an improvement of up to 10%.

In this paper, we propose a novel motion field estimation algorithm for which a Udisparity map and forward-and-backward error removal are applied in a vehicular environment. Generally, a motion exists in an image obtained by a camera attached to a vehicle by vehicle movement; however, the obtained motion vector is inaccurate because of the surrounding environmental factors such as the illumination changes and vehicles shaking. It is, therefore, difficult to extract an accurate motion vector, especially on the road surface, due to the similarity of the adjacent-pixel values; therefore, the proposed algorithm first removes the road surface region in the obtained image by using a U-disparity map, and uses then the optical flow that represents the motion vector of the object in the remaining part of the image. The algorithm also uses a forward–backward error-removal technique to improve the motion-vector accuracy and a vehicle’s movement is predicted through the application of the RANSAC (RANdom SAmple Consensus) to the previously obtained motion vectors, resulting in the generation of a motion field. Through experiment results, we show that the performance of the proposed algorithm is superior to that of an existing algorithm.

First-principles calculations have been used to investigate the effects of Al and Mg doping on the prevention of degradation phenomena in Li(Ni0.8Co0.1Mn0.1)O2 cathode materials. Specifically, we have examined the effects of dopants on the suppression of oxygen evolution and cation disordering, as well as their correlation. It is found that Al doping can suppress the formation of oxygen vacancies effectively, while Mg doping prevents the cation disordering behaviors, i.e., excess Ni and Li/Ni exchange, and Ni migration. This study also demonstrates that formation of oxygen vacancies can facilitate the construction of the cation disordering, and vice versa. Delithiation can increase the probabilities of formation of all defect types, especially oxygen vacancies. When oxygen vacancies are present, Ni can migrate to the Li site during delithiation. However, Al and Mg doping can inhibit Ni migration, even in structures with preformed oxygen defects. The analysis of atomic charge variations during delithiation demonstrates that the degree of oxidation behavior in oxygen atoms is alleviated in the case of Al doping, indicating the enhanced oxygen stability in this structure. In addition, changes in the lattice parameters during delithiation are suppressed in the Mg-doped structure, which suggests that Mg doping may improve the lattice stability.

Vacant parking-slot recognition is one of the core elements for achieving a fully automated parking assistant system. The recognition of a vacant parking-slot is typically executed in two stages: parking-slot recognition and slot-occupancy classification. In most previous works, considering the recognised slot results as a grid cell, the existence of objects is evaluated through the slot examination in slot-occupancy classification. Despite moderate performance, however, these methods cannot distinguish the parking-slot availability of moving or small objects and permit marks, since all previous methods only focus on the occupancy of typically parked vehicles. In this study, the authors present a camera-based available parking-slot recognition algorithm based on slot-context analysis. By utilising abundant visual features, they propose a new algorithm handling diverse available parking-slot conditions. For parking-slot recognition, they propose a new methodology of extracting and associating line-markings. The proposed algorithm includes a slot-validation step that identifies multiple slot contexts in a probabilistic integration and has more flexibility on irregular patterns. In the slot-occupancy classification stage, reliable parking-availability detection is achieved through visual-slot features including the histogram of gradient and frequency-magnitude features, via a support vector machine. Simulations and vehicle-level experiments demonstrate the robustness of the proposed algorithm in diverse conditions.

To realize high-speed intrusion detection by accommodating many regular expression (regex)-based signatures and growing network link capacities, we propose the Service TimE–Aware Load-balancing (STEAL) algorithm. This work is motivated from the observation that utilization of a many-core network intrusion detection system (NIDS) is influenced by unfair computational distribution among many-core NIDS nodes. To avoid such unfair computational distribution, STEAL is designed to dynamically distribute a large volume of traffic among manycore NIDS nodes based on packet service time, which is represented by the deep packet time in many-core NIDS nodes. From experiments, we show that compared to the commonly used loadbalancing algorithm based on arrival rate, STEAL increases the number of received packets (i.e., decreases the number of dropped packets) in many-core NIDS. Specifically, by integrating an open source NIDS (i.e. Bro) with STEAL, we show that even under attack-dominant traffic and with many signatures, STEAL can rapidly improve the performance of many-core NIDS to realize highspeed intrusion detection.

The Gaussian mixture probability hypothesis density (GM-PHD) filter has been widely adopted to track multiple targets, because it can effectively handle target birth/death without the track-to-measurement data association process. However, the GM-PHD filter is known to have serious problems related to birth intensity generation and target tractability. In addition, weight underestimation/overestimation may occur if there are missing detections or measurement clutters. Since these problems may lead to severe estimation errors, many researchers have tried to find solutions. However, none of the researchers have been successful at solving these problems simultaneously. In this paper, we propose a robust multitarget tracking scheme based on the GM-PHD filter to improve estimation accuracy, even if there are many false detections. The proposed scheme includes the processing step of evaluating multiple states/measurements, which is designed to overcome the weight underestimation/overestimation problems. Furthermore, it includes generating the birth intensity for the next iteration using measurements not associated with any tracked states. We also show that the proposed method can be extended to nonlinear Gaussian models. The simulation results demonstrate that the proposed scheme can provide relatively accurate multitarget estimates compared with the previous approaches when the measurements include many false positives/negatives.

The densification behavior and electrical conductivity of Ce0.8Gd0.2O1.9 ceramics with lithium oxide concentrations ranging from 0 to 7 mol% were investigated. The sintered density was found to increase with increasing Li2O content up to 2 mol% and then to decrease somewhat upon further Li20 addition. Dense Ce0.8Gd0.2O1.9 ceramics with 97% of the theoretical density could be obtained by sintering the milled mixture with 2 mol% Li2O addition at 1250 degrees C for 5 h. The conductivity of the 2 mol% Li2O-added specimen showed a maximum value of 4.99 x 10(-3) Ω(-1) x cm(-1) at 700 degrees C. Pure Ce0.8Gd0.2O1.9 ceramics needed to be sintered at 1550 degrees C in order to obtain an equivalent theoretical density and conductivity. The addition of Li2O was found to promote the sintering properties and electrical conductivities of Gd2O3-doped CeO2.

First-principles calculations and in-situ X-ray diffraction (XRD) experiment have been performed to investigate the structural changes and potential oxygen evolution in Li(Ni ( y ) Co (1−y)/2 Mn (1−y)/2 )O 2 (y = 0.33, 0.56, and 0.78) and LiNiO 2 cathode materials during electrochemical activities. They reveal that as the Ni content is increased, the change in lattice parameters becomes more significant during charging, which could induce the mechanical degradation in the cathode. Detailed analysis of delithiation process demonstrates the origin of the characteristic variations in the lattice parameters. Open-circuit voltage (OCV) calculations and charge–discharge experiments suggest that at the same voltage, more Li ions are extracted from Ni-rich structures, leading to further structural instability. Atomic charge analysis indicates that Ni mainly participates in charge compensation process in the practical range of delithiation, and then Co covers after further severe Li extraction. With increasing Ni content, the oxidation of oxygen, which is initially more oxidized, is accelerated during charging; this result explains the poorer oxygen stability observed in Ni-rich structures.

Organic/inorganic multilayer structures were fabricated for extreme flexibility as well as enhanced moisturebarrier property. The organic and inorganic layers for the structures were formed by plasma polymerization and atomic layer deposition, respectively. The layers were grown alternately to form the organic/inorganic multilayer structures on a plastic substrate. To accomplish extreme flexibility of the barriers, ultra-thin aluminum oxide layers were grown by the atomic layer deposition and sandwiched by a flexible plasma-polymer layer. The moisture-barrier films were then confirmed to retain the initial barrier property even after 10,000 times of bending at a radius as small as 3mm when the barrier structure was located at a neutral plane.

Highly accurate and robust real-time localization is an essential technique for various autonomous driving applications. Numerous localization methods have been proposed that combine various types of sensors, including an environmental sensor, IMU and GPS. However, the usage of a single environmental sensor is rather fragile. Although the use of multi-environment sensors is a better alternative, fusion methods from previous studies have not adequately compensated for shortcomings in dissimilar sensors or have not considered errors in the pre-built map. In this paper, we propose a decentralized localization framework using heterogeneous map-matching sources. Decentralized localization performs two independent map-matchings and integrates them with a stochastic situational analysis model. By applying a stochastic model, the reliability of the two map matchings is collected and system stability is verified. A number of experiments with autonomous vehicles within the actual driving environment have shown that combining multiple map-matching sources ensures more robust results than the use of a single environmental sensor.

While object tracking for 3D point clouds has been widely researched in recent years, most trackers employ a direct point-to-point matching method under the assumption that target object clouds are dense, although the method is not suitable for sparse point clouds. In this paper, we introduce a novel object-tracking strategy that enables even sparse point clouds to be tracked properly. The strategy involves estimating distributions, called as Estimation of Vertical Distributions (EVD), by the proposed interpolation method to augment data and by a point-to-distribution matching technique. The EVD step generates vertical distributions of unoccupied areas on a target object using the distributions of the occupied areas and then seeks the optimal solution through a coarse-to-fine grid search to guarantee real-time performance. In order to verify the proposed tracking algorithm, we have tested our tracker on real world data collected by our own platform, and the results have demonstrated that the tracker outperforms other trackers.

Because of rapid growth of data size, the number of data storage has been increased. When using multiple data storages, a distribute file system is essential to insure the availability of data files. The power consumption is a major problem when using a distributed file system with many data storages. Previous works have aimed at reducing the energy consumption with efficient file block layout by changing some data servers into stand-by mode. The file block migration has not been seriously considered because migration causes large cost. But when we consider addition of a new data server or file, file block migration is needed. This paper formulates the minimization of data block migration as an ILP optimization problem and solves it using branch-and-bound method. Using this technique, we can maximize the number of stand-by data servers with the minimum number of file block movement. However, computation time of branch-and-bound method of an ILP optimization problem increases exponentially as the problem size grows. Therefore this paper also proposes a data block and data server grouping method to solve many small ILP problems.

New zirconium oxide (ZrO2)-based organic-inorganic multilayers were fabricated and tested for flexible moisture barriers and compared with typical aluminum oxide (Al2O3)-based multilayers. We report that amorphous ZrO2 had a better intrinsic barrier property than that of amorphous Al2O3. Due to the lower elastic modulus of ZrO2, the ZrO2-based structures had better flexibility than that of the Al2O3-based structures. The ZrO2-based barrier was superior to the Al2O3-based barrier not only for flexibility but also for barrier performance. The barrier property and flexibility of the ZrO2-based structures were enhanced by about 20% and 30% over those of the Al2O3-based structures, respectively.

The effect of bi-functional coatings consisting of Zr and phosphate (P) on the electrochemical performance of Li1.0Ni0.8Co0.15Mn0.05O2 (NCM) has been investigated. The presence of various types of Zr and P compounds such as oxides (ZrO2 and Li2ZrO3) and phosphates (Zr2P2O9, ZrP2O7 and LiZr2(PO4)3) in the coating was confirmed by experiments as well as density functional theory (DFT) calculations. When the NCM samples were coated with the Zr/P hybrid material, the cycle retention and the amount of removed Li residuals (LiOH, Li2CO3) were enhanced by the synergistic effect from Zr and P. The NCM sample coated with a Zr/P layer with a Zr/P ratio of 1 : 1 exhibited an increase in the initial capacity (209.3 mA h g−1) compared to the pristine sample (207.4 mA h g−1) at 0.1C, owing to the formation of the coating layer. The capacity retention of the Zr/P coated sample (92.4% at the 50th cycle) was also improved compared to that of the pristine NCM sample (90.6% at the 50th cycle). Moreover, the amount of Li residuals in the Zr/P coated NCM sample was greatly reduced from 3693 ppm (pristine NCM) to 2525 ppm (Zr/P = 5 : 5).

In this paper, we propose novel motion field estimation method using U-disparity map and forward-backward error removal in vehicles environment. Generally, in an image obtained from a camera attached in a vehicle, a motion vector occurs according to the movement of the vehicle. but this motion vector is less accurate by effect of surrounding environment. In particular, it is difficult to extract an accurate motion vector because of adjacent pixels which are similar each other on the road surface. Therefore, proposed method removes road surface by using U-disparity map and performs optical flow about remaining portion. forward-backward error removal method is used to improve the accuracy of the motion vector. Finally, we predict motion of the vehicle by applying RANSAC(RANdom SAmple Consensus) from acquired motion vector and then generate motion field. Through experimental results, we show that the proposed algorithm performs better than old schemes.

Pedestrian detection has largely been researched as one of the important technologies for autonomous driving vehicle and preventing accidents. There are two categories for pedestrian detection, camera-based and LIDAR-based. LIDAR-based methods have the advantage of the wide angle of view and insensitivity of illuminance change while camera-based methods have not. However, there are several problems with 3D LIDAR, such as insufficient resolution to detect distant pedestrians and decrease in detection rate in a complex situation due to segmentation error and occlusion. In this paper, two methods using GM-PHD filter are proposed to improve the poor rates of pedestrian detection algorithms based on 3D LIDAR. First one improves detection performance and resolution of object by automatic accumulation of points in previous frames onto current objects. Second one additionally enhances the detection results by applying the GM-PHD filter which is modified in order to handle the poor situation to classified multi target. A quantitative evaluation with autonomously acquired road environment data shows the proposed methods highly increase the performance of existing pedestrian detection algorithms.

In vehicular communications, a group key that is only shared by the group subscribers is widely used for encrypting or signing multicast messages. A major issue with the group key is how to efficiently and securely transmit an updated group key for which numerous research results have been published. However, in the existing schemes, because vehicles receive the group key only from roadside units (RSUs), vehicles that are out of communication range cannot receive the updated group key. In this case, those vehicles can receive no data nor participate in group communication until they get the newest group key from the RSU. To overcome the aforementioned out-of-range limitation, we propose a new group key delivery scheme based on vehicle-to-vehicle (V2V) communication. The proposed scheme enables the vehicle in a group out of the communication range of the RSU to receive the group key from other vehicles in the same group. Two main parts of the scheme are a vehicle-verification protocol and a group-key-verification protocol. With the first protocol, a group key sender can verify if the vehicle that tries to receive the group key is a group member, and with the second protocol, the group key receiver can check the integrity of the group key. The two protocols include no interaction with a third party and, moreover, do not use any unencrypted privacy-sensitive information within a vehicle. By using mathematical analysis, we theoretically demonstrate that our scheme is safe against possible attacks. In addition, by using simulations, we show that the V2V group key delivery scheme guarantees faster group key dissemination than the existing schemes.

When the structural analysis is performed for the concrete lining of the water pressure tunnel, many parameters are considered such as relaxed ground loads, internal water pressure, external water pressure, the shrinkage of the concrete lining, grouting pressure, etc. But, there are no standards and manuals for the structural analysis for the concrete lining of the water pressure tunnel. Above all, the external water pressure has an much effect on the stability of tunnel. So, in case that permeability of ground is large, the external water pressure should be decreased by installation of weep hole, or reinforced ground by ground improvement grouting should be pressed by the external water pressure instead. But, when weep hole is installed to reduce the external water pressure, the many problems may me occurred. Thus, reasonable approach for treatment of the external water pressure is necessary if weep hole is not installed. Therefore, the purpose of this study is to analyze design cases and studies for treatment of the external water pressure in performing structural analysis for the concrete lining of the water pressure tunnel, and to find reasonable method for tunnel lining modeling which is the treatment of the external water pressure according to permeability of ground and consequently the design of ground improvement grouting.Keywords: Water pressure tunnel, Weep hole, Concrete lining modeling, External water pressure, Internal water pressure

Safe and efficient deoxygenation of benzylic alcohols was accomplished using a droplet-based microreactor, which induced rapid mixing of the hydrophilic reagents (HI and H 3 PO 2 ) in the aqueous phase and the alcohols in the organic phase. We investigated the relationship between the surface-to-volume ratio of the organic droplets and the reaction rate, and examined the effect of temperature and concentration of HI and H 3 PO 2 in the reaction. The deoxygenated product was separated from the strongly acidic aqueous solution, and HI in the aqueous solution could be recycled. The reaction afforded 99% yield at 120 °C with HI (3.4 M HI) and H 3 PO 2 (4.8 M).

Photooxygenation of monoterpenes was conducted in two continuous flow reactors. The first, suitable for lab-scale research, had a maximum yield of 99.9%, and the second, focused on industrial applications, showed a daily output that was 270.0-fold higher than that in batch systems. The use of sunlight instead of an LED lamp gave 68.28% conversion.

Ni-rich layered oxides are considered to be a promising cathode material with high capacity, and their surface structure should be extensively explored to understand the complex associated phenomena. We investigated the surface stability and morphology of LiNiO

Ni-rich layered oxides are promising cathode materials due to their high capacities. However, their synthesis process retains a large amount of Li residue on the surface, which is a main source of gas generation during operation of the battery. In this study, combined with simulation and experiment, we propose the optimal metal phosphate coating materials for removing residual Li from the surface of the Ni-rich layered oxide cathode material LiNi0.91Co0.06Mn0.03O2. First-principles-based screening process for 16 metal phosphates is performed to identify an ideal coating material that is highly reactive to Li2O. By constructing the phase diagram, we obtain the equilibrium phases from the reaction of coating materials and Li2O, based on a database using a DFT hybrid functional. Experimental verification for this approach is accomplished with Mn3(PO4)2, Co3(PO4)2, Fe3(PO4)2, and TiPO4. The Li-removing capabilities of these materials are comparable to the calculated results. In addition, electrochemical performances up to 50 charge/discharge cycles show that Mn-, Co-, Fe-phosphate materials are superior to an uncoated sample in terms of preventing capacity fading behavior, while TiPO4 shows poor initial capacity and rapid reduction of capacity during cycling. Finally, Li-containing equilibrium phases examined from XRD analysis are in agreement with the simulation results.

In driving environments, other vehicles are one of the most frequently appearing close range objects from the ego-vehicle. Thus, the development of high accuracy vehicle recognition algorithms is essential for safe and efficient automated driving. However, detecting vehicles with consistently high accuracy is difficult because there are various vehicle types with different appearances, such as sedans, buses, trucks, and SUVs. This intra-class variation must be addressed or, irregular recognition performance can occur, depending on vehicle type. Conventional machine learning-based algorithms are inadequate to address this problem because they are mostly trained on samples of entire appearance. Considering the wide variability in vehicle appearance, collecting samples of every vehicle type may not be ideal. In this study, we propose a vehicle recognition algorithm using common appearance characteristics of every vehicle type. Rectangular shapes are captured by a 3D LIDAR while tires and bumpers are captured by a monocular camera. Angular features extracted from these common appearances are then fused by the Dempster-Shafer theory framework for vehicle recognition. By performing real-world experiments, we demonstrated that common appearances captured by the proposed algorithm provide sufficiently generalized features to recognize diverse vehicle types in urban driving environments.

Over the last decade, automated vehicles have been widely researched and their massive potential has been verified through several milestone demonstrations. However, there are still many challenges ahead. One of the biggest challenges is integrating them into urban environments in which dilemmas occur frequently. Conventional automated driving strategies make automated vehicles foolish in dilemmas such as making lane-change in heavy traffic, handling a yellow traffic light and crossing a double-yellow line to pass an illegally parked car. In this paper, we introduce a novel automated driving strategy that allows automated vehicles to tackle these dilemmas. The key insight behind our automated driving strategy is that expert drivers understand human interactions on the road and comply with mutually-accepted rules, which are learned from countless experiences. In order to teach the driving strategy of expert drivers to automated vehicles, we propose a general learning framework based on maximum entropy inverse reinforcement learning and Gaussian process. Experiments are conducted on a 5.2 km-long campus road at Seoul National University and demonstrate that our framework performs comparably to expert drivers in planning trajectories to handle various dilemmas.

The solvent evaporation method on the structural changes and surface chemistry of the cathode and the effect of electrochemical performance of Li1.0Ni0.8Co0.15Mn0.05O2 (NCM) has been investigated. After dissolving of Li residuals using minimum content of solvent in order to minimize the damage of pristine material and the evaporation time, the solvent was evaporated without filtering and remaining powder was re-heated at 700 °C in oxygen environment. Two kinds of solvent, de-ionized water and diluted nitric acid, were used as a solvent. The almost 40% of Li residuals were removed using solvent evaporation method. The NCM sample after solvent evaporation process exhibited an increase in the initial capacity (214.3 mAh/g) compared to the pristine sample (207.4 mAh/g) at 0.1C because of enhancement of electric conductivity caused by decline of Li residuals. The capacity retention of NCM sample after solvent evaporation process (96.0% at the 50th cycle) was also improved compared to that of the pristine NCM sample (90.6% at the 50th cycle). The uniform Li residual layer after solvent treated and heat treatment acted like a coating layer, leading to enhance the cycle performance. The NCM sample using diluted nitric acid showed better performance than that using de-ionized water.

Ultra-thin aluminum-oxide layers were grown by atomic layer deposition, and plasma-polymer layers derived from an n-hexane precursor were prepared by plasma polymerization. Hybrid nanolaminates were fabricated using one-cycle-grown aluminum-oxide layers and 50 nm-thick plasma polymer layers, and their moisture-barrier property was measured by an electrical calcium test. The moisture-barrier property of the hybrid nanolaminates was exponentially enhanced as the number of dyads increased, indicating that an ultra-thin single-cycle-grown aluminum oxide worked as a good moisture barrier. A 20-dyad hybrid nanolaminate composed of two-cycle-grown aluminum-oxide layers and 50 nm-thick plasma-polymer layers shows a water vapor transmission rate of 1×10 −3 g/m 2 ·day.

Batteries mounted on electric vehicles (EVs) are often damaged by high peak power and rapid charging/discharging cycles, which are originated from repetitive acceleration/deceleration of vehicles particularly in urban situations. To reduce battery damage, the battery/supercapacitor (SC) hybrid energy storage system (HESS) has been considered as a solution because the SC can act as a buffer against large magnitudes and rapid fluctuations in power. While the traditional purpose of employing the HESS in EVs is to minimize the magnitude/variation of battery power or power loss, the previous approaches proposed for controlling the HESS have some drawbacks; they neither consider these objectives simultaneously nor reflect real-time load dynamics for computing the SC reference voltage. In this paper, we present a power control framework consisting of two stages: one for computing the SC reference voltage and another for optimizing the power flowing through the HESS. In the presented framework, we propose a methodology for calculating the SC reference voltage considering the real-time load dynamics without given future operation profiles. In addition, we formulate the HESS power control problem as a convex optimization problem that minimizes the magnitude/fluctuation of battery power and power loss at the same time. The optimization problem is formulated so that it can be repeatedly solved by general solvers in polynomial time. Simulation results carried out on MATLAB show that the magnitude/variation of battery power and power loss can be concurrently reduced in real time by the proposed framework.

The kinetic theory for the growth of pearlite in binary and ternary steels is implemented to ensure local equilibrium at the transformation front with austenite, while accounting for both boundary and volume diffusion of solutes. Good agreement is on the whole observed with published experimental data, although the reported growth rate at the lowest of temperatures is much smaller than predicted. To investigate this, experiments were conducted to replicate the published data. It is found that the cooperation between cementite and ferrite breaks down at these temperatures, and surface relief experiments are reported to verify that the resulting transformation product is not bainite.

The densification behavior and electrical conductivity of Ce0.8Y0.2O1.9 ceramics with gallia concentrations ranging from 0 to 5 mol. % were investigated. The sintered density was found to increase with increasing Ga2O3 content up to 1 mol. % and then to decrease upon further Ga2O3 addition. Dense Ce0.8Y0.2O1.9 ceramics with 94% of the theoretical density could be obtained by sintering the milled mixture with 1 mol. % Ga2O3 addition at 1400°C for 5 h. The conductivity of the 1 mol. % Ga2O3-added specimen showed a maximum value of 1.37 × 10−2 Ω−1· cm−1 at 700°C. Pure Ce0.8Y0.2O1.9 ceramics needed to be sintered at 1550°C in order to obtain an equivalent theoretical density and conductivity. The introduction of Ga2O3 doping had a good effect on the sintering properties and electrical conductivities of Y2O3-doped CeO2.

The prospect of enhancing the hardness of low alloy steel for the manufacture of fasteners is examined using ausforming, in which the austenite is deformed rapidly at a low temperature to increase its dislocation density before quenching in order to obtain the harder martensite. Surprisingly, small deformations accomplish large gains in hardness and the dislocation density of martensite, with diminishing returns at larger deformations. The main contribution to the hardness has been identified as the extra dislocations inherited by the martensite from the deformed austenite, rather than the refinement of microstructure by the ausforming process. Clear evidence is reported for the mechanical stabilisation of the austenite due to ausforming. Tempering heat treatments tend to diminish the advantages of ausforming.