Brain-like Computing Based on Distributed Representations and Neurodynamics

A key to overcoming the limitations of classical artificial intelligence and to deal well with enormous amounts of information might be brain-like computing in which distributed representations of information are processed by dynamical systems without using symbols. We present a method for such computing. We constructed an inference system using a nonmonotone neural network, which is a kind of recurrent neural network with continuous-time dynamics. This system deduces a conclusion according to state transitions of the network in which knowledge is embedded as trajectory attractors. It has the powerful ability of analogical reasoning without special treatment for exceptional knowledge. We also propose a method of linking different neurodynamical systems and show that two mutually interacting systems can process complex spatiotemporal patterns.     

Fabrication of X-rays mask with carbon membrane for diffraction gratings

We have produced diffraction gratings for obtaining high resolution X-ray phase imaging, such as X-ray Talbot interferometer. These diffraction gratings were required to have a fine, high accuracy, high aspect ratio structure. Therefore, we decided to use the X-rays lithography technique that used synchrotron radiation of the directivity for a manufacture process. The accuracy of the completed structure depends largely on the accuracy of the X-ray mask. In our group, a resin material is conventionally used for the membrane of large X-ray masks. However, X-ray masks comprising a resin membrane have the disadvantage that, after several cycles of X-ray exposure, they crease and sag due to X-ray-derived heat. As a substitute for the conventional resin membrane, we experimentally fabricated a new X-ray mask using a carbon wafer membrane. The newly fabricated X-ray mask was subjected to X-ray exposure experiment. We succeeded in making the structure body which was almost shape. And the experimental results verified that the new mask did not deteriorate even when used repeatedly, demonstrating that it was highly durable.     

Fabrication of high hardness Ni mold with electroless nickel–boron thin layer

The nickel electroforming method using a high-concentration nickel sulfamate bath is commonly used to fabricate micro metal molds in the LIGA process; however, this method does not produce micro metal molds of sufficient hardness. One means of improving the hardness of micro metal molds made using the nickel electroforming method is to include additives in the nickel plating solution. Another method is nickel alloy plating or a similar technique. In this research, we used a nickel–boron (Ni–B) electroless alloy plating method to obtain a hard nickel plated film having hardness of 832 Hv. It was also ascertained that Ni–B electroless alloy plated film retains its high hardness even during heat treatment in conditions of 250°C for 1 h. To deal with the high stresses developed in high-hardness plated films, we proposed double-layer nickel electroforming. This method is covered and used on conventional nickel electroforming layer by high hardness micro mold. High hardness micro metal mold using double-layer was fabricated by nickel electroforming and Ni–B electroless alloy plating method.     

Universality Issues in Reversible Computing Systems and Cellular Automata (Extended Abstract)

In this survey, we deal with the problem how a universal computer can be constructed in a reversible environment. We discuss this problem based on the frameworks of reversible Turing machines, reversible logic circuits, and reversible cellular automata. We can see that in spite of the constraint of reversibility, there are several very simple reversible systems that have universal computing ability.     

Trends in highly scalable crossbar-based packet switch architecture

A combined input and crosspoint queued (CICQ) switch is receiving significant attention to be the next generation high speed packet switch for its scalability; however, a multi-cabinet implementation of a combined input and crosspoint queued (CICQ) switch unavoidably introduces a large round-trip time (RTT) latency between the line cards and switch fabric, resulting a large crosspoint (CP) buffer requirement. In this paper, virtual crosspoint queues (VCQs) that significantly reduces the CP buffer requirement of the CICQ switch is investigated. The VCQs unit resides inside the switch fabric, is dynamically shared among virtual output queues (VOQ) from the same source port, and is operated at the line rate, making the implementation practical. A threshold-based exhaustive round-robin (T-ERR) arbitration is employed to reduce buffer hogging at VCQ. The T-ERR at VCQ and CP arbiters serves packets residing in a longer queue more frequently than packet residing in a shorter queue. Consequently, the T-ERR, drastically increases the throughput of the CICQ switch with small CP buffers. A multi-cabinet implementation of CICQ switch do not support multicasting traffic well since a combination of small CP buffer in the switch fabric and a large RTT latency between the line cards and switch fabric results in non-work conservation of the intra-switch link. Deployment of multicast FIFO buffer between the input buffer and CP buffer shows a promise. With its ability to achieve high throughput independent of RTT and switch port size, the integration of the VCQ architecture and T-ERR scheduler to the CICQ switch is ideal for supporting ever-increasing Internet traffic that requires higher data rate, larger switch size, and efficient multicasting.     

Development of ‘UR-System 2’: a training system for recovery of motor function of plegic upper limb after stroke

We developed a new training system, called ‘UR-System 2’, for the recovery of motor function of a plegic upper limb after stroke. UR-System 2, which was redesigned based on knowledge gained from clinical tests using UR-System 1, was newly equipped with a function for testing the motor function of patients, such as isolated movement. In order to verify the effectiveness of UR-System 2, we conducted the following procedures: we verified the safety of training in healthy persons; we gathered standard data on pronation and supination of healthy forearms during tests; and, we verified that the test function was effective in evaluating the recovery effect of isolated movement in two hemiplegic patients.     

Multi-contact vertical ladder climbing with an HRP-2 humanoid

We describe the research and the integration methods we developed to make the HRP-2 humanoid robot climb vertical industrial-norm ladders. We use our multi-contact planner and multi-objective closed-loop control formulated as a QP (quadratic program). First, a set of contacts to climb the ladder is planned off-line (automatically or by the user). These contacts are provided as an input for a finite state machine. The latter builds supplementary tasks that account for geometric uncertainties and specific grasps procedures to be added to the QP controller. The latter provides instant desired states in terms of joint accelerations and contact forces to be tracked by the embedded low-level motor controllers. Our trials revealed that hardware changes are necessary, and parts of software must be made more robust. Yet, we confirmed that HRP-2 has the kinematic and power capabilities to climb real industrial ladders, such as those found in nuclear power plants and large scale manufacturing factories (e.g. aircraft, shipyard) and construction sites.     

Molecular dynamics simulations of diffusion of submonolayer polar liquid lubricant films on solid surfaces

A fundamental understanding of the diffusion phenomena of submonolayer polar liquid films is important for achieving reliable lubrication between moving mechanical parts separated by a nanometer-sized gap. To acquire this understanding, we conducted molecular dynamics (MD) simulations of diffusion phenomena of submonolayer polar perfluoropolyether (PFPE) Zdol films on solid surfaces. To improve the accuracy of these simulations, we developed an all-atom model that includes hydrogen-bond potential and refined atomic charges for Zdol molecules and tested it through MD simulations of spreading of step-shaped submonolayer PFPE films. Our MD simulations reproduced the experimentally observed effects of polar end groups on the diffusion speed and molecular conformation of Zdol. We then conducted MD simulations of self-diffusion of submonolayer Zdol films; these simulations demonstrated that as the thickness of the submonolayer Zdol films decreases, molecular conformation becomes flatter and the self-diffusion coefficient decreases. These changes in molecular conformation partially explain our experimental finding that the spreading of step-shaped submonolayer polar PFPE films slows down with decreasing initial thickness.     

A design of high optical efficiency wedge type light guide plate adopting twin inclined cylindrical surface

We propose wedge type light guide plate (LGP) adopting twin inclined cylindrical surfaces (TICS) at the incident part of a LGP. It has higher optical efficiency than conventional wedge type LGP. In dimension of LGP which the thickness of incident part is 1.5 mm and that of luminance part is 0.7 mm, the optical efficiency of wedge type LGP adopting TICS is expected to be 96% in comparison with a normal flat LGP. In addition, there is an advantage to reduce the non‐uniformity of vicinity of light emitted diodes (LEDs) by TICS reflecting light laterally. However, it is necessary that the LED centers are aligned with the intersections of TICS. The misalignment of LED causes non‐uniformity of optical profile that is slightly larger than a normal LGP.