Showing papers by "Toyota published in 2017"

Deep Neural Network for Structural Prediction and Lane Detection in Traffic Scene

Abstract: Hierarchical neural networks have been shown to be effective in learning representative image features and recognizing object classes. However, most existing networks combine the low/middle level cues for classification without accounting for any spatial structures. For applications such as understanding a scene, how the visual cues are spatially distributed in an image becomes essential for successful analysis. This paper extends the framework of deep neural networks by accounting for the structural cues in the visual signals. In particular, two kinds of neural networks have been proposed. First, we develop a multitask deep convolutional network, which simultaneously detects the presence of the target and the geometric attributes (location and orientation) of the target with respect to the region of interest. Second, a recurrent neuron layer is adopted for structured visual detection. The recurrent neurons can deal with the spatial distribution of visible cues belonging to an object whose shape or structure is difficult to explicitly define. Both the networks are demonstrated by the practical task of detecting lane boundaries in traffic scenes. The multitask convolutional neural network provides auxiliary geometric information to help the subsequent modeling of the given lane structures. The recurrent neural network automatically detects lane boundaries, including those areas containing no marks, without any explicit prior knowledge or secondary modeling.

All-printed thin-film transistors from networks of liquid-exfoliated nanosheets.

Abstract: All-printed transistors consisting of interconnected networks of various types of two-dimensional nanosheets are an important goal in nanoscience. Using electrolytic gating, we demonstrate all-printed, vertically stacked transistors with graphene source, drain, and gate electrodes, a transition metal dichalcogenide channel, and a boron nitride (BN) separator, all formed from nanosheet networks. The BN network contains an ionic liquid within its porous interior that allows electrolytic gating in a solid-like structure. Nanosheet network channels display on:off ratios of up to 600, transconductances exceeding 5 millisiemens, and mobilities of >0.1 square centimeters per volt per second. Unusually, the on-currents scaled with network thickness and volumetric capacitance. In contrast to other devices with comparable mobility, large capacitances, while hindering switching speeds, allow these devices to carry higher currents at relatively low drive voltages.

The renaissance of hydrides as energy materials

Abstract: Materials based on hydrides have been the linchpin in the development of several practical energy storage technologies, of which the most prominent example is nickel–metal hydride batteries. Motivated by the need to meet the future's energy demand, the past decade has witnessed substantial advancements in the research and development of hydrides as media for hydrogen energy storage. More recently, new and rapidly evolving discoveries have positioned hydrides as highly promising materials for future electrochemical energy storage, such as electrolytes for mono- and divalent batteries, and anodes for lithium-ion batteries. In addition, the potential of hydrides in efficient power transmission has been recently revealed. In this Review, we highlight key advances and illustrate how the versatility of hydrides has not only yielded a meaningful past, but also ensures a very bright future. Discoveries of new hydride properties beyond those expected are ushering in a new era in hydride research and development. This Review covers these rapidly evolving advancements; explains their relevance to future energy storage and transmission applications; and proposes future research directions.

Fervent Hype behind Magnesium Batteries: An Open Call to Synthetic Chemists—Electrolytes and Cathodes Needed

Abstract: Magnesium metal is a superior anode which has double the volumetric capacity of lithium metal and has a negative reduction potential of −2.37 V vs. the standard hydrogen electrode. A major benefit of magnesium is the apparent lack of dendrite formation during charging which is one of the crucial concerns of using a lithium metal anode. In this Review, we highlight the foremost research in the development of electrolytes and cathodes and discuss some of the significant challenges which must be overcome in realizing a practical magnesium battery.

Predicting Catalytic Activity of Nanoparticles by a DFT-Aided Machine-Learning Algorithm

Abstract: Catalytic activities are often dominated by a few specific surface sites, and designing active sites is the key to realize high-performance heterogeneous catalysts. The great triumphs of modern surface science lead to reproduce catalytic reaction rates by modeling the arrangement of surface atoms with well-defined single-crystal surfaces. However, this method has limitations in the case for highly inhomogeneous atomic configurations such as on alloy nanoparticles with atomic-scale defects, where the arrangement cannot be decomposed into single crystals. Here, we propose a universal machine-learning scheme using a local similarity kernel, which allows interrogation of catalytic activities based on local atomic configurations. We then apply it to direct NO decomposition on RhAu alloy nanoparticles. The proposed method can efficiently predict energetics of catalytic reactions on nanoparticles using DFT data on single crystals, and its combination with kinetic analysis can provide detailed information on stru...

Unveiling What Is Written in the Stars: Analyzing Explicit, Implicit, and Discourse Patterns of Sentiment in Social Media

Abstract: Deciphering consumer’s sentiment expressions from Big Data (e.g., online reviews) has become a managerial priority to monitor product and service evaluations. However, Sentiment Analysis, the process of automatically distilling sentiment from text, provides little insight regarding the language granularities beyond the use of positive and negative words. Drawing on Speech Act Theory, this study provides a fine-grained analysis of the implicit and explicit language used by consumers to express sentiment in text. An empirical text mining study using more than 45,000 consumer reviews, demonstrates the differential impacts of activation levels (e.g., tentative language), implicit sentiment expressions (e.g., commissive language), and discourse patterns (e.g., incoherence) on overall consumer sentiment (i.e., star ratings). In two follow-up studies, we demonstrate that these speech act features also influence the readers’ behavior and are generalizable to other social media contexts such as Twitter and Facebook. We contribute to research on consumer sentiment analysis by offering a more nuanced understanding of consumer sentiments and their implications

Multipolicy decision-making for autonomous driving via changepoint-based behavior prediction: Theory and experiment

Abstract: This paper reports on an integrated inference and decision-making approach for autonomous driving that models vehicle behavior for both our vehicle and nearby vehicles as a discrete set of closed-loop policies. Each policy captures a distinct high-level behavior and intention, such as driving along a lane or turning at an intersection. We first employ Bayesian changepoint detection on the observed history of nearby cars to estimate the distribution over potential policies that each nearby car might be executing. We then sample policy assignments from these distributions to obtain high-likelihood actions for each participating vehicle, and perform closed-loop forward simulation to predict the outcome for each sampled policy assignment. After evaluating these predicted outcomes, we execute the policy with the maximum expected reward value. We validate behavioral prediction and decision-making using simulated and real-world experiments.

The DARPA Robotics Challenge Finals: Results and Perspectives

Abstract: The DARPA Robotics Challenge DRC program conducted a series of prize-based competition events to develop and demonstrate technology for disaster response. This article provides the official and definitive account of DRC Finals as the culmination of the DRC program. The article details the eight tasks Drive, Egress, Door, Valve, Wall, Surprise [Plug and Switch], Rubble [Obstacle or Debris], and Stairs constituting the Challenge, and describes how the competition encouraged supervised autonomous operation by intentionally degrading the communications channel between the remote human operators. The article presents the results of the DRC Finals and places those results in perspective by identifying both strengths and weaknesses of robot performance exhibited at the competition.

Are All-Solid-State Lithium-Ion Batteries Really Safe?-Verification by Differential Scanning Calorimetry with an All-Inclusive Microcell.

Abstract: Although all-solid-state lithium-ion batteries (ALIBs) have been believed as the ultimate safe battery, their true character has been an enigma so far. In this paper, we developed an all-inclusive-microcell (AIM) for differential scanning calorimetry (DSC) analysis to clarify the degree of safety (DOS) of ALIBs. Here AIM possesses all the battery components to work as a battery by itself, and DOS is determined by the total heat generation ratio (ΔH) of ALIB compared with the conventional LIB. When DOS = 100%, the safety of ALIB is exactly the same as that of LIB; when DOS = 0%, ALIB reaches the ultimate safety. We investigated two types of LIB-AIM and three types of ALIB-AIM. Surprisingly, all the ALIBs exhibit one or two exothermic peaks above 250 °C with 20–30% of DOS. The exothermic peak is attributed to the reaction between the released oxygen from the positive electrode and the Li metal in the negative electrode. Hence, ALIBs are found to be flammable as in the case of LIBs. We also attempted to impr...

Ultrahigh thermoelectric power factor in flexible hybrid inorganic-organic superlattice.

Abstract: Hybrid inorganic-organic superlattice with an electron-transmitting but phonon-blocking structure has emerged as a promising flexible thin film thermoelectric material. However, the substantial challenge in optimizing carrier concentration without disrupting the superlattice structure prevents further improvement of the thermoelectric performance. Here we demonstrate a strategy for carrier optimization in a hybrid inorganic-organic superlattice of TiS2[tetrabutylammonium] x [hexylammonium] y , where the organic layers are composed of a random mixture of tetrabutylammonium and hexylammonium molecules. By vacuum heating the hybrid materials at an intermediate temperature, the hexylammonium molecules with a lower boiling point are selectively de-intercalated, which reduces the electron density due to the requirement of electroneutrality. The tetrabutylammonium molecules with a higher boiling point remain to support and stabilize the superlattice structure. The carrier concentration can thus be effectively reduced, resulting in a remarkably high power factor of 904 µW m-1 K-2 at 300 K for flexible thermoelectrics, approaching the values achieved in conventional inorganic semiconductors.

Simultaneous Contact, Gait and Motion Planning for Robust Multi-Legged Locomotion via Mixed-Integer Convex Optimization

Abstract: Traditional motion planning approaches for multilegged locomotion divide the problem into several stages, such as contact search and trajectory generation. However, reasoning about contacts and motions simultaneously is crucial for the generation of complex whole-body behaviors. Currently, coupling theses problems has required either the assumption of a fixed gait sequence and flat terrain condition, or nonconvex optimization with intractable computation time. In this letter, we propose a mixed-integer convex formulation to plan simultaneously contact locations, gait transitions, and motion, in a computationally efficient fashion. In contrast to previous works, our approach is not limited to flat terrain nor to a prespecified gait sequence. Instead, we incorporate the friction cone stability margin, approximate the robot's torque limits, and plan the gait using mixed-integer convex constraints. We experimentally validated our approach on the HyQ robot by traversing different challenging terrains, where nonconvexity and flat terrain assumptions might lead to suboptimal or unstable plans. Our method increases the motion robustness while keeping a low computation time.

Robust Online Monitoring of Signal Temporal Logic

Abstract: Signal temporal logic (STL) is a formalism used to rigorously specify requirements of cyberphysical systems (CPS), i.e., systems mixing digital or discrete components in interaction with a continuous environment or analog components. STL is naturally equipped with a quantitative semantics which can be used for various purposes: from assessing the robustness of a specification to guiding searches over the input and parameter space with the goal of falsifying the given property over system behaviors. Algorithms have been proposed and implemented for offline computation of such quantitative semantics, but only few methods exist for an online setting, where one would want to monitor the satisfaction of a formula during simulation. In this paper, we formalize a semantics for robust online monitoring of partial traces, i.e., traces for which there might not be enough data to decide the Boolean satisfaction (and to compute its quantitative counterpart). We propose an efficient algorithm to compute it and demonstrate its usage on two large scale real-world case studies coming from the automotive domain and from CPS education in a Massively Open Online Course setting. We show that savings in computationally expensive simulations far outweigh any overheads incurred by an online approach.

Study of Electrochemical Phenomena Observed at the Mg Metal/Electrolyte Interface

Abstract: The interface between Mg metal and electrolyte is a key factor affecting Mg battery performance. Switchable interfacial phenomena, involving apparent surface electrochemical inhibition under open-circuit voltage and reactivation upon electrochemical polarization, were investigated with various Mg electrolyte systems, under both electrochemically static and dynamic conditions. Most notably, it was found that such behavior appears to be unique for the Mg system, implying that correct control of the interface is of considerable practical concern in Mg batteries. This new challenge must be addressed in order to achieve high-energy and high-durability rechargeable Mg batteries.

Dynamic Modulation of Radiative Heat Transfer beyond the Blackbody Limit.

Abstract: Dynamic control of electromagnetic heat transfer without changing mechanical configuration opens possibilities in intelligent thermal management in nanoscale systems. We confirmed by experiment that the radiative heat transfer is dynamically modulated beyond the blackbody limit. The near-field electromagnetic heat exchange mediated by phonon–polariton is controlled by the metal–insulator transition of tungsten-doped vanadium dioxide. The functionalized heat flux is transferred over an area of 1.6 cm2 across a 370 nm gap, which is maintained by the microfabricated spacers and applied pressure. The uniformity of the gap is validated by optical interferometry, and the measured heat transfer is well modeled as the sum of the radiative and the parasitic conductive components. The presented methodology to form a nanometric gap with functional heat flux paves the way to the smart thermal management in various scenes ranging from highly integrated systems to macroscopic apparatus.

Development of the fuel cell vehicle mirai

Abstract: The use of electricity and hydrogen is considered a promising way to help resolve environmental and energy-related issues. Based on this concept, fuel cell (FC) vehicles (FCVs) have been developed with the aim of achieving their widespread adoption in the future. This article describes the technical challenge to be overcome in the actual production of FCs installed in FCV Mirai. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

A solution-processed TiS2/organic hybrid superlattice film towards flexible thermoelectric devices

Abstract: Liquid-exfoliation has proven to be a scalable and versatile technique to produce high-yield two-dimensional nanosheets in graphene, BN, layered perovskites and transition metal dichalcogenides. This also provides new insights into the construction of novel nanoelectronics and nanophotonics through the assembly of nanosheets. Here we present a simple exfoliation-and-reassembly approach to produce a flexible n-type TiS2/organic hybrid film for low-temperature thermoelectric applications. The obtained film shows a superlattice structure with alternative layers of TiS2 and organic molecules. Charge transfer occurs when TiS2 and organic molecules form intercalation complexes, which gives rise to a high electrical conductivity but a low Seebeck coefficient. However, the power factor can be further enhanced by annealing the film under vacuum, and the value reaches 210 μW m−1 K−2 at room temperature in this study. Our flexible thermoelectric device can generate a high power density of 2.5 W m−2 at a temperature gradient of 70 K, which hits a new record among organic-based thermoelectric devices.

Inverse Multipath Fingerprinting for Millimeter Wave V2I Beam Alignment

Abstract: Efficient beam alignment is a crucial component in millimeter wave systems with analog beamforming, especially in fast-changing vehicular settings. This paper proposes a position-aided approach where the vehicle's position (e.g., available via GPS) is used to query the multipath fingerprint database, which provides prior knowledge of potential pointing directions for reliable beam alignment. The approach is the inverse of fingerprinting localization, where the measured multipath signature is compared to the fingerprint database to retrieve the most likely position. The power loss probability is introduced as a metric to quantify misalignment accuracy and is used for optimizing candidate beam selection. Two candidate beam selection methods are developed, where one is a heuristic while the other minimizes the misalignment probability. The proposed beam alignment is evaluated using realistic channels generated from a commercial ray-tracing simulator. Using the generated channels, an extensive investigation is provided, which includes the required measurement sample size to build an effective fingerprint, the impact of measurement noise, the sensitivity to changes in traffic density, and beam alignment overhead comparison with IEEE 802.11ad as the baseline. Using the concept of beam coherence time, which is the duration between two consecutive beam alignments, and parameters of IEEE 802.11ad, the overhead is compared in the mobility context. The results show that while the proposed approach provides increasing rates with larger antenna arrays, IEEE 802.11ad has decreasing rates due to the larger beam training overhead that eats up a large portion of the beam coherence time, which becomes shorter with increasing mobility.

Fast Trajectory Optimization for Agile Quadrotor Maneuvers with a Cable-Suspended Payload

Abstract: Reference EPFL-CONF-232508 URL: http://rpg.ifi.uzh.ch/docs/RSS17_Foehn.pdf Record created on 2017-11-21, modified on 2017-11-21

Procedural Generation of Videos to Train Deep Action Recognition Networks

Abstract: Deep learning for human action recognition in videos is making significant progress, but is slowed down by its dependency on expensive manual labeling of large video collections. In this work, we investigate the generation of synthetic training data for action recognition, as it has recently shown promising results for a variety of other computer vision tasks. We propose an interpretable parametric generative model of human action videos that relies on procedural generation and other computer graphics techniques of modern game engines. We generate a diverse, realistic, and physically plausible dataset of human action videos, called PHAV for Procedural Human Action Videos. It contains a total of 39,982 videos, with more than 1,000 examples for each action of 35 categories. Our approach is not limited to existing motion capture sequences, and we procedurally define 14 synthetic actions. We introduce a deep multi-task representation learning architecture to mix synthetic and real videos, even if the action categories differ. Our experiments on the UCF101 and HMDB51 benchmarks suggest that combining our large set of synthetic videos with small real-world datasets can boost recognition performance, significantly outperforming fine-tuning state-of-the-art unsupervised generative models of videos.

On the Reliability of Half-Cell Tests for Monovalent (Li+, Na+) and Divalent (Mg2+, Ca2+) Cation Based Batteries

Abstract: A comprehensive study is reported entailing a comparison of Li, Na, K, Mg, and Ca based electrolytes and an investigation of the reliability of electrochemical tests using half-cells. Ionic conductivity, viscosity, and Raman spectroscopy results point to the cationsolvent interaction to follow the polarizing power of the cations, i.e. Mg2+ > Ca2+ > Li+ > Na+ > K+ and to divalent cation based electrolytes having stronger tendency to form ion pairs - lowering the cation accessibility and mobility. Both increased temperature and the use of anions with delocalized negative charge, such as TFSI, are effective in mitigating this issue. Another factor impeding the divalent cations mobility is the larger solvation shells, as compared to those of monovalent cations, that in conjunction with stronger solvent - cation interactions contribute to slower charge transfer and ultimately a large impedance of Mg and Ca electrodes. An important consequence is the non-reliability of the pseudo-reference electrodes as these present both significant potential shifts as well as unstable behaviors. Finally, experimental protocols in order to achieve consistent results when using half-cell set-ups are proposed.

Cu2ZnSnS4 photovoltaic cell with improved efficiency fabricated by high-temperature annealing after CdS buffer-layer deposition

Abstract: To improve the photovoltaic properties of Cu2ZnSnS4 (CZTS) cells, we investigated the effect of both the thickness of the deposited CdS layers and the post-annealing temperature following CdS deposition on the photovoltaic properties of CZTS cells using a two-layer CZTS structure. By depositing a thin CdS layer (40 nm) followed by high temperature annealing (603 K), we observed a remarkable increase in the short-circuit current density because of the enhancement of the external quantum efficiency in the wavelength range of 400–800 nm. The best CZTS cell exhibited a conversion efficiency of 9.4% in the active area (9.1% in the designated area). In addition, we also fabricated a CZTS cell with open-circuit voltage of 0.80 V by appropriately tuning the composition of the CZTS layers. Copyright © 2016 John Wiley & Sons, Ltd.

The role of iodide in the formation of lithium hydroxide in lithium–oxygen batteries

Abstract: Lithium iodide has been studied extensively as a redox-mediator to reduce the charging overpotential of Li–oxygen (Li–O2) batteries. Ambiguities exist regarding the influence of lithium iodide on the reaction product chemistry and performance of lithium–oxygen batteries. In this work, we examined the role of lithium iodide on the reduction product chemistry under two conditions: (i) mixing KO2 with lithium salts and (ii) discharging Li–oxygen batteries at high and low overpotentials, in the presence of an ether-based electrolyte with different ratios of H2O : LiI. The addition of iodide to electrolytes containing water was found to promote the formation of LiOOH·H2O, LiOH·H2O and LiOH at the expense of Li2O2. At low H2O : LiI ratios (lower than 5), LiOH instead of Li2O2 was formed, which was accompanied by the oxidation of iodide to triodide while at high H2O : LiI ratios (12, 24, 134), a mixture of Li2O2, LiOOH·H2O and LiOH·H2O was observed and no triiodide was detected. The reaction between peroxide Li2O2 and/or superoxide LiO2 with H2O to form LiOH is facilitated by increased water acidity by strong I−–H2O interactions as revealed by 1H NMR and FT-IR measurements. This mechanism of LiOH formation in the presence of LiI and H2O was also found upon Li–O2 cell discharge, which is critical to consider when developing LiI as a redox mediator for Li–O2 batteries.

Significant increase of oral bacteria in the early pregnancy period in Japanese women.

Abstract: Aim Oral microflora during pregnancy is critical to oral health care in the mother and her child. We examined the changes in the oral microbiota between pregnancy and nonpregnancy periods. Methods The study was performed using 132 healthy pregnant women enrolled from Hiroshima City Asa Citizens Hospital and 51 healthy nonpregnant women as control. During pregnancy, 132 subjects were assessed for seven microbial species by the cultured method and polymerase chain reaction at the early (7–16 weeks gestation), the middle (17–28 weeks), and the late (29–39 weeks) pregnancy periods. Pregnant women completed a series of questionnaires regarding oral and systemic health and lifestyle habits. Results The total cultivable microbial counts in the early pregnancy were significantly higher than that of the nonpregnant women (P < 0.05). The incidences of Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans in gingival sulcus during the early and middle pregnancy were significantly higher than the nonpregnant group (P < 0.05), while Prevotella intermedia and Fusobacterium nucleatum did not change. Candida species were more frequently detected during the middle and late pregnancy. Conclusion The data suggest that pregnancy, especially in the early periods, promotes the proliferation of microorganisms in the oral cavity and facilitates a colonization of periodontal pathogens.

A learning approach for real-time temporal scene flow estimation from LIDAR data

Abstract: Many autonomous systems require the ability to perceive and understand motion in a dynamic environment. We present a novel algorithm that estimates this motion from raw LIDAR data in real-time without the need for segmentation or model-based tracking. The sensor data is first used to construct an occupancy grid. The foreground is then extracted via a learned background filter. Using the filtered occupancy grid, raw scene flow between successive scans is computed. Finally, we incorporate these measurements in a filtering framework to estimate temporal scene flow. We evaluate our method on the KITTI dataset.

Chiral Superstructure Mesophases of Achiral Bent-Shaped Molecules – Hierarchical Chirality Amplification and Physical Properties

Abstract: Chiral mesophases in achiral bent-shaped molecules have attracted particular attention since their discovery in the middle 1990s, not only because of their homochirality and polarity, but also due to their unique physical/physicochemical properties. Here, the most intriguing results in the studies of such symmetry-broken states, mainly helical-nanofilament (HNF) and dark-conglomerate (DC) phases, are reviewed. Firstly, basic information on the typical appearance and optical activity in these phases is introduced. In the following section, the formation of mesoscopic chiral superstructures in the HNF and DC phases is discussed in terms of hierarchical chirality. Nanoscale phase segregation in mixture systems and gelation ability in the HNF phase are also described. In addition, some other related chiral phases of bent-shaped molecules are shown. Recent attempts to control such mesoscopic chiral structure and the alignment/confinement of HNFs are also discussed, along with several examples of their fascinating advanced physical properties, i.e. huge enhancement of circular dichroism, electro- and photo-tunable optical activities, chirality-induced nonlinear optics (second-harmonic-generation circular difference and electrogyration effect), enhanced hydrophobicity through the dual-scale surface morphological modulation, and photoconductivity in the HNF/fullerene binary system. Future prospects from basic science and application viewpoints are also indicated in the concluding section.