Nissan Research Center

Yokosuka, Japan

Nissan Research Center

Yokosuka, Japan
Time filter
Source Type

Ahmad T.,University of Nevada, Reno | Ilstrup D.,Nissan Research Center | Emami E.,University of Nevada, Reno | Bebis G.,University of Nevada, Reno
IEEE Intelligent Vehicles Symposium, Proceedings | Year: 2017

This paper investigates the use of Convolutional Neural Networks for classification of painted symbolic road markings. Previous work on road marking recognition is mostly based on either template matching or on classical feature extraction followed by classifier training which is not always effective and based on feature engineering. However, with the rise of deep neural networks and their success in ADAS systems, it is natural to investigate the suitability of CNN for road marking recognition. Unlike others, our focus is solely on road marking recognition and not detection; which has been extensively explored and conventionally based on MSER feature extraction of the IPM images. We train five different CNN architectures with variable number of convolution/max-pooling and fully connected layers, and different resolution of road mark patches. We use a publicly available road marking data set and incorporate data augmentation to enhance the size of this data set which is required for training deep nets. The augmented data set is randomly partitioned in 70% and 30% for training and testing. The best CNN network results in an average recognition rate of 99.05% for 10 classes of road markings on the test set. © 2017 IEEE.

News Article | August 25, 2016

NASA's Ames Research Center is in the second year of a five-year partnership with Nissan to develop an autonomous electric car, targeted for release in 2020. That collaboration includes adapting software from Ames' K-10 and K-REX robotic rovers, which were designed to test concepts for future missions such as laying elements for a giant radio telescope on the moon's far side. Terry Fong, director of the Intelligent Robotics Group at Ames, leads the NASA team in the self-driving car project. His counterpart is Maarten Sierhuis, director of the Nissan Research Center, Silicon Valley, which is about an eight-minute drive (manual or autonomous) from Ames. Before coming to Nissan in 2013, Sierhuis spent 12 years as a NASA engineer, working in the same Intelligent Systems Division at Ames as Fong. "Anybody who works at NASA for a long time, you get kind of indoctrinated to believe that there is no problem difficult enough that you cannot solve it," Sierhuis said. They have their work cut out for them in perfecting a self-driving car suited to real-world challenges. "Frankly, it's pretty easy to address the 80 or 90 percent of things that you see on the road all the time," Fong said. "What's really hard is addressing the anomalies, things that only pop up once in a while." Fong's daughter came of driving age around the same time the NASA-Nissan collaboration began. His experience in teaching her to drive illustrates one of the chief difficulties in teaching a car to drive itself. "A tree had fallen halfway across the road, blocking our side of the road," Fong said. "My daughter said, 'what do I do?' And I said, 'well, let's pull over, look around carefully, and if it's clear, then it's okay for you to drive on the wrong side of the road for a little bit to get around this.' How do you develop enough intelligence in a self-driving car for the car to make that decision, to recognize that it's okay to break the rules of the road for a short period of time? I frankly have no idea." Aside from equipping the car to deal with unusual events, Fong said the toughest part of the task is enabling it to understand and communicate clearly with human drivers, who sometimes forgo turn signals or drive erratically. "When we are still in this world of some cars that are manually driven and some that are autonomous, we are going to have to figure out how we signal between the two," he said, "until we flip all the way over where everything is autonomous and then we don't have to worry about that anymore." "I have a group of anthropologists here that are studying how people behave in traffic," Sierhuis said. "We use social science to help us understand what people do in the activities of driving, bicycling, being a pedestrian. You need to understand this in order to build a car that can live and behave as a social team member in a society where people move around." Autonomous cars could free people to spend their commuting time in activities more productive or enjoyable than stewing in traffic jams (see sidebar). Evidence is emerging that robots can drive more safely than people as well. And the environment, too, has a rooting interest in self-driving cars. It's easy to see that automated cars can be programmed to drive energy-efficiently, free of jackrabbit starts and seatbelt-straining stops. Less obvious is that a system of self-driving cars could increase the capacity of existing highways, reducing the need to build more. "Right now, when you are driving on a high-speed freeway, you have to maintain a two-to-four-second gap between the cars, because that's what we humans recognize as being kind of safe in terms of reaction time," Fong said. "But there's no physical reason why the cars couldn't be closer and all come to a stop safely, especially if they do it in a coordinated fashion. Then you could double or triple the number of cars on the road at any given time. And if they are all working really well, you can also do that at a higher speed." Combining intelligent cars with intelligent traffic-management systems can boost efficiency even further. Sierhuis described using his native Netherlands as a test bed. "In the region of North Holland, which includes Amsterdam and other big cities, every single traffic light, every bridge is connected," he said. "The traffic management system knows what's going on. It has all the traffic flow data. "We have demonstrated that our autonomous vehicle software can communicate bidirectionally with the traffic lights. It can predict whether a traffic light will be green as much as two miles before the car gets to that light. And the vehicle can ask the traffic light for increased green time so it won't have to stop. If it can't get that green time, it might reroute itself. It's all about making this an efficient system." "NASA has long had a very safety-oriented culture," Fong said. "We develop complex systems that we know are operating in difficult environments, so we try to manage the risk and safety associated with operating them. In working with car companies, we've been trying to apply the same kind of approaches." Sierhuis agrees, saying that the attitude he brings to his work at Nissan has been transplanted from his work at NASA. "We were developing an intelligent agent for spacesuits," Sierhuis said, recounting his days as a NASA engineer. He recalls an astronaut telling him, "Maarten, just remember, my spacesuit is my work environment, my home and my life support system. Don't screw it up. If you screw it up, I'm dead." "So I'm thinking, OK, this car is a robot," Sierhuis said. "It's a freaking heavy robot and it drives 65, 70 miles per hour on the highway and a person is inside of it. And then I think, it's my home, my work environment, my life support. So this is really like the astronaut talking about his spacesuit. We'd better not screw it up." Maarten Sierhuis' first car as a young man in the Netherlands was a replica of the car Roger Moore drove in the James Bond movie, For Your Eyes Only: a yellow Citroën 007 Special Edition with bullethole decals."It was the coolest car ever," Sierhuis said. "Except the problem was, there were only a handful in the Netherlands and only one in Rotterdam and that was mine. So the police knew me. Every Friday and Saturday night, it was really hard driving that car home after partying because everybody knew who it was. "I do enjoy driving," he said. "But 95 percent of the driving I do, I want to give up. Because 95 percent of the driving I do is commuting from home to work. Driving is kind of a necessary evil in the Bay Area, especially during office hours." Terry Fong concurs. "Every morning when I get on Highway 85 and it's bumper-to-bumper, I would love to be able to sit back and spend the time catching up on my email instead of just watching the car ahead of me." Upcoming self-driving cars may not have the ejector seats or oil-slick sprayers of a James Bond car. But they will give people a license to relax and enjoy the ride. Explore further: Stay calm and drive on—tips to manage road rage

Nozaki H.,Kogakuin University | Makita M.,Nissan Research Center | Yoshino T.,Kogakuin University
SAE Technical Papers | Year: 2011

The relation of the front wheel steering angle to the steering wheel angle in electric vehicles is changing due to the "steer-by-wire" method, which is based on an electric signal. With this method, excellent maneuverability is possible in various driving situations. Therefore, this steer control method technique is considered in this study. It was clarified that steer-bywire requires an improvement in the control stability in emergency maneuvers and the delay of counter steering in drift cornering without causing a sense of driver incompatibility. (Here, the sense of incompatibility was defined as feeling by which the harmony between the steer intention of the driver and the vehicle movement was lost.) (Here, the drift cornering shows cornering done in the area with counter steering where the rear wheel exceeded the maximum cornering force.) One control stability method is Proportional Derivative (PD) steering assistance, which is dependent on the anticipated driving situations. In addition, the negative camber angle control that is proportional to the steer angle can improve lateral acceleration in critical cornering. These results clarify that three-dimensional control of the wheel posture angle according to a specific driving situation greatly improves driving stability. Copyright © 2011 SAE International.

Zhu D.,Tsinghua University | Nanbu T.,Nissan Research Center | Ren N.,Northwestern University | Yasuda Y.,Nissan Research Center | Wang Q.J.,Northwestern University
Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology | Year: 2010

A model-based virtual texturing approach has been developed and applied to design, generate, 'test', and evaluate textured surfaces through numerical simulations. A series of studies on the numerical generation and performance evaluation of textured surfaces in a lubricated concentrated contact has been conducted, which includes (a) numerical generation of a large variety of textures considering possible geometric imperfections that exist in reality due to tooling design considerations and fabrication errors; (b) determination of texture depth, size, and area density; (c) texture distribution pattern selection; (d) bottom shape comparison and design optimization; (e) investigation of the influence of surface relative motion; (f) prediction of performance deviation caused by texture shape imperfections; and (g) evaluation of the effect of originally machined roughness. The present study was conducted using the deterministic mixed elastohydrodynamic lubrication (EHL) model recently developed, modified, and validated, which appears to be useful for surface texture design based on comparative performance evaluation in a wide range of operating conditions.

News Article | February 15, 2017

Nissan has a new tool to get safe autonomous cars on the road quickly: Humans. While advances in AI are making vehicles smarter, we are not yet at the point where autonomous vehicles can completely understand how to handle unpredictable driving situations, Nissan said in a press release. To more quickly realize the potential of autonomous driving, Nissan CEO and chairman of the board Carlos Ghosn announced a new technology called Seamless Autonomous Mobility (SAM), in a keynote address at the 2017 Consumer Electronics Show last week. SAM pairs in-vehicle artificial intelligence (AI) with support from a human call center, allowing human operators to remotely control an autonomous car when the vehicle finds itself in an unfamiliar or dangerous driving situation, such as a construction zone or an accident. SAM was born from Nissan's partnership with NASA on autonomous car research and development. It uses a version of NASA's Visual Environment for Remote Virtual Exploration (VERVE) software, which was used to navigate the rovers on Mars. SEE: CES 2017 Special Coverage (CNET) | CES 2017 (TechRepublic) | CES 2017: The Big Trends for Business (ZDNet) For example, the press release stated, imagine an autonomous car is driving through city streets, and comes across an accident, with police using hand signals to direct traffic across double yellow lines and against traffic lights. While the vehicle sensors can tell where the obstacles are, human judgement is still needed to understand the situation and determine the best course of action. That's where the call center comes in. Using SAM, the car will be smart enough to bring itself to a safe stop and ping the command center. "The request is routed to the first available mobility manager—a person who uses vehicle images and sensor data (streamed over the wireless network) to assess the situation, decide on the correct action, and create a safe path around the obstruction," the press release stated. "The mobility manager does this by 'painting' a virtual lane for the vehicle to drive itself through." Once the car is out of the accident area and back on its route, it resumes fully autonomous operations. Nissan's AI also learns from the experience: Other Nissans in the area can communicate with SAM, and as the system learns the proper path, it shares that information with those other vehicles. "Our goal is to change the transportation infrastructure," said Maarten Sierhuis, former NASA scientist and director of the Nissan Research Center in Silicon Valley, in the press release. "We want to reduce fatalities and ease congestion. We need a huge number of vehicles out there. What we are doing at Nissan is finding a way so that we can have this future transportation system not in 20 years or more, but now." SAM's goal is to "use the human intelligence more strategically to support a larger system of autonomous mobility—and to help improve the artificial intelligence of the vehicles in real-time." Not to mention, it quashes at least some fears of humans losing their jobs to robots. The announcements are part of the Nissan Intelligent Mobility blueprint, aimed at "transforming how cars are driven, powered, and integrated into wider society," according to the press release. "We invite others to join us, as well, from tech partners to e-commerce companies, ride-hailing and car-sharing platforms, and social entrepreneurs who can help us to test and develop new vehicles and services, and make sure everyone has access to the latest technologies and services that bring value to their lives," said Ghosn in the press release. At CES, Ghosn also announced that Nissan and Japanese internet company DeNA will begin testing autonomous cars for commercial services this year in Japan. He also unveiled plans to launch a new Nissan LEAF with ProPILOT technology to enable autonomous driving functionality for single-lane highway driving. The manufacturer will also continue its partnership with Microsoft on connected cars with Cortana services.

Ragoisha G.A.,Belarusian State University | Osipovich N.P.,Belarusian State University | Bondarenko A.S.,University of Twente | Zhang J.,Nissan Research Center | And 2 more authors.
Journal of Solid State Electrochemistry | Year: 2010

Multi-frequency ac responses of Pt in aqueous solutions of sulphuric and perchloric acids have been characterised in cycles of Pt oxide anodic formation and cathodic reduction as functions of electrode potential, using the two orders frequency range below 1 kHz, where the double layer responded jointly with Faradaic processes. The potentiodynamic impedance spectra were fitted to an equivalent circuit, which contained double-layer capacitance in parallel with charge transfer resistance, R, and constant phase element (CPE). Double-layer capacitance has shown minima (20-25 μF cm -2 in 0.5 M H 2SO 4 and 25-35 μF cm -2 in 1 M HClO 4) in the double-layer region and more than twofold increase in the platinum oxidation region. R -1 and CPE have shown maxima in the regions of platinum oxidation and reduction. Potentiodynamic curves of raw impedance data were also self-descriptive in monitoring platinum oxidation and reduction in both solutions and in presence of chloride. © 2008 Springer-Verlag.

Kato S.,Tokyo Institute of Technology | Watanabe Y.,Tokyo Institute of Technology | Kurokawa Y.,Tokyo Institute of Technology | Yamada A.,Tokyo Institute of Technology | And 3 more authors.
Japanese Journal of Applied Physics | Year: 2012

30-nm-diameter silica nanoparticles with a carboxyl radical (COO -) were successfully dispersed on an amino-treated silicon wafer at about 20 nm intervals owing to the repulsion among nanoparticles with negative charges. The dispersed silica nanoparticles were used as the mask for the preparation of silicon nanowire (SiNW) arrays by metal-assisted chemical etching (MAE). The diameter of the prepared SiNWs was approximately 30nm from their transmission electron microscope image. © 2012 The Japan Society of Applied Physics.

Ohishi M.,Tokyo Institute of Technology | Ono Y.,Nissan Research Center | Fushinobu K.,Tokyo Institute of Technology
Nanoscale and Microscale Thermophysical Engineering | Year: 2013

Transport resistance of reactant gases in the membrane of polymer electrolyte fuel cells (PEFCs) was investigated by measuring the mass transport amount of gas species with microprobes inserted along the membrane thickness direction. The diffusion resistance inside the membrane and the interfacial transport resistance including dissolution resistance were separately evaluated by combining experiments and a theoretical model. It was found that diffusion resistance decreased with an increase in gas humidity and temperature. The measurement technique using the microprobe is a useful tool for understanding the gas transport phenomena in the membrane. Copyright © Taylor & Francis Group, LLC.

Strmcnik D.,Argonne National Laboratory | Uchimura M.,Argonne National Laboratory | Uchimura M.,Nissan Research Center | Wang C.,Argonne National Laboratory | And 6 more authors.
Nature Chemistry | Year: 2013

The development of hydrogen-based energy sources as viable alternatives to fossil-fuel technologies has revolutionized clean energy production using fuel cells. However, to date, the slow rate of the hydrogen oxidation reaction (HOR) in alkaline environments has hindered advances in alkaline fuel cell systems. Here, we address this by studying the trends in the activity of the HOR in alkaline environments. We demonstrate that it can be enhanced more than fivefold compared to state-of-the-art platinum catalysts. The maximum activity is found for materials (Ir and Pt 0.1 Ru 0.9) with an optimal balance between the active sites that are required for the adsorption/dissociation of H 2 and for the adsorption of hydroxyl species (OHad). We propose that the more oxophilic sites on Ir (defects) and PtRu material (Ru atoms) electrodes facilitate the adsorption of OHad species. Those then react with the hydrogen intermediates (Had) that are adsorbed on more noble surface sites. © 2013 Macmillan Publishers Limited. All rights reserved.

Takahashi I.,Nissan Research Center | Kocha S.S.,Nissan Motor Co.
Journal of Power Sources | Year: 2010

Widespread research in the field of fuel cells necessitates easily verifiable and reproducible benchmarks for characterizing properties such as electrochemical area (ECA), oxygen reduction reaction (ORR) specific and mass activity (is, im) as well as durability of electrocatalysts. Ex situ characterization of electrocatalysts deposited as thin-film rotating disk electrodes (TF-RDE) in liquid electrolytes as well as in their original dry powder state has been conducted. Commercially available Pt on carbon support (Pt/C) catalyst serving as a baseline benchmark and heat treated Pt/C and Pt-alloy/C catalysts were investigated as examples of higher activity and durability materials. A detailed description of the preparation and optimization of catalyst inks, measurement protocols, and analysis of ORR kinetic parameters and durability rates are provided to form a basis for consistent screening and benchmarking of new and improved catalysts for proton exchange membrane fuel cells (PEMFCs). Preparation of highly-dispersed ink slurries formulated using various water-isopropanol mixtures and deposited as TF-RDEs were demonstrated to significantly affect the magnitude of measured ECA and activity. The ECA, is and im for the baseline Pt/C were determined to be 100 m2 g-1, 292 μA cm-2Pt and 266 mA mg-1Pt in 0.1 M HClO4 at 25 °C and 10 mV s-1. Strong adsorption of anions on Pt/C in sulfuric acid was shown to have a deleterious effect on its activity and durability. Related ORR kinetic parameters such as the activation energy (ΔH = 38 kcal mol-1) as well as the experimental reaction order (m ∼ 0.75) with respect to oxygen were determined to provide a basis for converting literature results to a common pressure and temperature. © 2010 Elsevier B.V. All rights reserved.

Loading Nissan Research Center collaborators
Loading Nissan Research Center collaborators