Search

A hardware CPG model for biped gait control using pulse-type hardware neural networks(P-HNNs) composed of analog electronic circuits to generate time-series neural signal patterns for controlling human walking and running is reported. Human walking and running are controlled by central pattern generators located in a spinal cord. Apart from this spinal system, gait and gait speed is known to be changed by a signal from a higher center input to a brainstem-spinal cord projection. However, the specific circuit configuration of the neural circuits for controlling gait locomotion is still unknown. In recent studies, walking and running have been estimated to be controlled with five simple time-series pulse patterns generated by the spinal cord and the duration of their pulses. We have previously used P-HNNs to generate time-series neural signal pulse patterns for controlling human walking and running. In the model, the order of the time series was determined manually by trigger input. In addition, the change in the pulse pattern period when switching from walking to running was not reproduced. In this paper, we propose a hardware CPG model for biped gait control, with a network configuration capable of automatically determining the order of the time series, and changing the pulse pattern period. Based on the circuit configuration of the proposed CPG model, circuit simulations and CPG circuits fabricated with surface mount components were measured. The measured results confirm the automatic output of the neural signal patterns for controlling walking and running in time-series order. We also confirmed the pulse pattern period becomes shorter when switching from walking to running.

For constructing an isolated tooth identification system using deep learning, Igarashi et al. (2021) began constructing a learning model as basic research to identify the left and right mandibular first and second premolars. These teeth were chosen for analysis because they are difficult to identify from one another. The learning method itself was proven appropriate but presented low accuracy. Therefore, further improvement in the learning data should increase the accuracy of the model. The study objectives were to modify the learning data and increase the learning model accuracy for enabling the identification of isolated lower premolars.Static images of the occlusal surface of the premolars made from the dental plaster casts of dental students were used as the training, validation, and test data. A convolutional neural network with 32 hidden layers, AlexNet, convolutional architecture for fast feature embedding, and stochastic gradient descent was used to construct four learning models.The accuracy of the identification model increased using static images of the occlusal surface of the teeth with the adjacent teeth deleted as the training and validation data; however, a learning model that could perfectly identify the teeth could not be realized.Static images of the occlusal surface of the teeth with the adjacent teeth deleted should be used as both training and validation data. The ratio of the numbers of training, validation, and test data should be optimized.

The authors aim to develop an insect-type microrobot system which can operate autonomously like as an insect by mounting a sensor, a power supply, a motor, and a controller in a millimeter-sized body. Realization of robots the same as the performance of organisms is difficult with current robotics. In particular, small size actuator with low power consumption, which can actuate by small-sized energy source are difficult to realize. In this paper, the authors will discuss an electrostatic linear motor to move the legs of the microrobot with low energy consumption using a small power source.

Wetting is one of the omnipresent phenomena governed via natural laws. Moreover, surface wettability at non-ambient temperature especially at high temperature (30°C to 90°C) is of great importance in many industrial processes. In this study, Si wafers with various structures were fabricated to investigate wettability at different temperatures. Three shapes with micro-pillar structured surfaces were designed and fabricated. Pillar-structured surfaces were fabricated by photolithography and ICP etching. The temperature-dependent wettability of single-phase regime droplets was characterized using contact angle measurements. The wetting behavior of a water droplet was observed.

The authors are studying a robot controlling system using hardware neural networks (HNN). Previously, the authors succeeded to locomote the quadruped-type and the hexapod-type microrobot systems using the HNN integrated circuit (IC). The HNN IC outputs four-phase pulse waveforms to drive both microrobot systems. However, gait patterns need 1.0 s pulse period which requires large capacitors. A large capacitor could not mount on IC; therefore, the external capacitors mounted on the circuit board. The further miniaturization of the HNN IC needs to minimize the capacitors. In this paper, the authors constructed HNN which can generate a large pulse period without using large capacitors. The interstitial cell model is used as a basic element of the HNN to generate a large time constant. The authors designed two types of HNN which can generate a tripod gait pattern and a ripple gait pattern which are typical walking patterns of insects. As a result, designed HNN can generate gait patterns without using external capacitors. The capacitors could construct inside the IC using the interstitial cell model.

The authors are studying Hardware Neural Networks (HNN) for generating a driving waveform of a millimeter size insect-type microrobot. HNN can generate the pulse waveform such as the neural networks of living organisms. In the previous research, the HNN constructed by an Integrated Circuit (IC) could generate only a tripod gait pattern which is necessary to perform the locomotion of the microrobot. The microrobot can move six legs independently; thus, the mechanical structure of the microrobot allows the several gaits by changing the driving waveform. The HNN with switchable gait pattern could perform the different locomotion of the microrobot. In this paper, the authors discuss HNN which can switch the gait pattern of the microrobot. HNN can generate two gait pattern such as the tripod gait pattern and the ripple gait pattern which is typical gaits of insects. The tripod is for fast walking and the ripple is for slow walking, respectively. Usually, six cell body models mutually connected by 18 inhibitory synaptic models were required to generate the tripod gait pattern. Also, six cell body models mutually connected by 30 inhibitory synaptic models were required to generate the ripple gait pattern. The authors simplified the network with proposal excitatory-inhibitory switchable synaptic model. The HNN can simplify as six cell body models connected by two excitatory-inhibitory switchable synaptic model and ten inhibitory synaptic models. In addition, two types of gait patterns can be switched using a single external voltage source. As a result, The HNN generates tripod gait pattern with an external voltage of 3.0 V and ripple gait pattern with an external voltage of -3.0 V.

In this paper, miniature electromagnetic induction type MEMS generators are developed and demonstrate a power generation. The fabricated MEMS generators use a miniature ball-bearing for stable rotational motion for an air turbine. Moreover, the MEMS turbine is combined with the multilayer ceramic magnetic circuit because it realizes the miniature three-dimensional coil structure with the magnetic core. Two type magnetic circuits; a single-phase type and a three-phase type are designed. The maximum output power of the single-phase type generator was 2.41 mVA when the rotational speed was 290,135 rpm and the load resistance of 8ohm was connected. The maximum output power of the three-phase type generator was 3.01 mVA when the rotational speed was 228,484 rpm and the load resistance of 3 ohm was connected. To achieve more high output power, optimization of the magnetic flux introducing part shape is required to suppress the magnetic flux leakage. For the three-phase type generator, alignment error and difference of the coil inductance influence the output value.

This paper proposes a three-phase type MEMS air turbine generator. The developed generator is combined with a silicon air turbine and a ceramic magnetic circuit. The magnetic circuit for an electromagnetic induction type generator is fabricated by the multilayer ceramic technology. The fabricated multilayer circuit realizes a three-dimensional coil structure inside the monolithic ceramic structure. Moreover, the circuit shows low internal DC resistance because Ni-Cu-Zn ferrite and silver conductor are used for the magnetic material and the coil pattern. The miniature air turbine is fabricated by the MEMS process. To achieve the high rotational speed, the miniature ball-bearing is used, and the rim-shape rotor is used. In this generator, the distance between the magnet and the magnetic circuit is important factor. The magnet is held by the shaft through the ball-bearing and rim-type rotor, and it showed narrow distance to the magnetic circuit. Dimensions of the fabricated MEMS generator were 6.17 mm, 5.62 mm and 7.12 mm, length, width and height. The maximum rotational speed of the fabricated electromagnetic induction type MEMS air turbine generator was 228,484 rpm, when the flow rate was 2.9 l/min and the pressure was 0.3 MPa. The maximum three-phase output power was 3.01 mVA at load resistance of 3 ohm. The achieved value was 42 % of the theoretical power. The reason is alignment error of the coil pattern. To achieve more high output power, the alignment pattern will be introducing to suppress the assembled error.

The ideal microrobots are millimeter sized with integrated actuators, power sources, sensors, and controllers. Many researchers take the inspiration from insects for the mechanical or electrical designs to construct small sized robotic systems. Previously, the authors proposed and demonstrated microrobots which can replicate the tripod gait locomotion of an ant and the legs were actuated by shape memory alloy actuators. Shape memory alloy provided a large deformation and a large force, but the power consumption was as high as 94 mW to actuate a single leg. This paper discusses the silicon electrostatic inchworm motor chip with low energy consumption for the robot leg by using a small-size power source. The inchworm motor chip has actuated by electrostatic motors. The power consumption is low as 1.0 mW compering with shape memory alloy actuators. The reciprocal motion of the inchworm motor chip is power by the silicon photovoltaic cells. Results show the 7.5 mm square size photovoltaic cells could produce 60 V to actuate the inchworm motor chip. The generated force is enough to move the leg of the microrobot. We have shown actuation of the microrobot leg using an electrostatic inchworm motor chip. This result is the first instance of an electrostatic motor driving an off-chip structure.

Living organisms can act autonomously because biological neural networks process the environmental information in continuous time. Therefore, living organisms have inspired many applications of autonomous control to small-sized robots. In this chapter, a small-sized robot is controlled by a hardware artificial neural network (ANN) without software programs. Previously, the authors constructed a multilegged walking robot. The link mechanism of the limbs was designed to reduce the number of actuators. The current paper describes the basic characteristics of hardware ANNs that generate the gait for multilegged robots. The pulses emitted by the hardware ANN generate oscillating patterns of electrical activity. The pulse-type hardware ANN model has the basic features of a class II neuron model, which behaves like a resonator. Thus, gait generation by the hardware ANNs mimics the synchronization phenomena in biological neural networks. Consequently, our constructed hardware ANNs can generate multilegged robot gaits without requiring software programs.

This chapter explains how the MEMS microrobot system could perform the walking behavior of ants. MEMS microrobot system consists of micro-mechanical systems and micro-electro systems. The micro-mechanical systems mimic the alternating tripod gait of an ant by the shape memory alloy-type rotary actuator and the link mechanism. The micro-electro systems mimic the electrical activity of biological neural networks using the artificial neural networks IC. The artificial neural networks IC generates the driving pulses of shape memory alloy-type rotary actuator without using software programs. The micro-mechanical systems and micro-electro systems are integrated as a robot system. As a result, the authors show that the MEMS microrobot system could perform the ant-like walking behavior with a speed of 20 mm/min. The MEMS microrobot system was 0.079 g in weight, 4 mm width, 4 mm length, and 5 mm height in size. The robot system needs only the electrical power source as an external device.

This paper proposed a MEMS air turbine generator. The conventional miniature power sources like a battery required a replacement. Therefore, it is limitation for an IoT (Internet of Things) device, and it is required to develop a miniature generator. The proposed generator shows a millimetre scale structure that combined the different technology. The MEMS process is used for the miniature air turbine structure. To achieve the high output power, the electromagnetic induction type is employed. The miniature magnetic circuit is fabricated by the multilayer ceramic technology. Fluorocarbon gas is used as a working fluid to generate the rotational motion of the rotor. By this reason, the miniature power generation system will be realized in the future. Dimensions of the fabricated MEMS power generator were 5.62 mm, 6.17 mm, and 7.12 mm, respectively. The rotational motion was demonstrated using R245fa of fluorocarbon gas. The maximum rotational speed was 154,620rpm when the outside temperature was 46.3 Celsius degree. The output voltage was 10 mV at the load resistance of 1 kΩ. The calculated output power was 0.6 mVA at when the load resistance was 2 Ω.

This paper proposes miniaturized electromagnetic type power generators that compose of an air turbine by the MEMS (Micro Electro Mechanical Systems) technology and a magnetic circuit by the multilayer ceramic technology. Small and high energy density power supplies have been required with the increase in power consumption of electronic equipment. However, conventional power supplies have been reaching the theoretical limit. Accordingly, miniaturized power generators to replace the secondary battery have been studied. In this paper, two types of MEMS air turbine generators are proposed. The generators are the single-phase type MEMS air turbine generator and three-phase type, respectively. The generation capability of the single phase magnetic circuit was preliminary evaluated by a rotating permanent magnet which was attached to a spindle machine. When the rotational speed was 380000 rpm and a load resistance was 1 ohm, the output power reached to 1.47 mVA. Combined with MEMS air turbine, the generator showed maximum rotational speed of 30,000rpm, and the output power was 1.74 micro VA at load resistance of 1 ohm. The magnetic circuit of three phase type was evaluated by the spindle machine. When the rotational speed was 380000 rpm and a load resistance of 1 ohm, the output power was 3.3 mVA. Also, maximum rotational speed of the rotor of the MEMS air turbine generator was 18,000rpm, and the output power was 1.41 micro VA.

This paper proposes a multilayer ceramic coil for a wireless power transfer system that fabricated by a photo resist film process. The wireless power transfer technology has been focused as power supply of miniaturized communication devices because it is not required the connecting cable, and it is suitable to carry. Moreover, the electromagnetic induction type that is usually used is constructed by only coil structure for receiver. Therefore, it will be possible to conserve space inside the devices. Conventionally, the spiral structure coil is used for the transfer coil. However, the spiral coil is including problems such as a high internal resistance and consumption of large area. Moreover, the coil requires the magnetic material to catch the magnetic flux. Therefore, in this paper, the multilayer ceramic coil for the wireless power transfer system was proposed. The coil was fabricated by the multilayer ceramic technology which is used for miniaturized electric components. This technology has the advantage that it is possible to forming a helical structure and to use a magnetic material. In addition, the coil pattern is formed by the photo resist film process. In the conventional process, deformation of the conductive pattern by screen printing method was problem. The proposed process can be solved this problem because the conductive pattern was held by the resist film. In this paper, the rectangle conductive pattern in the cross-sectional was achieved. Moreover, the cross-sectional area was changed by the thickness of the resist film, and the coil pattern that has the cross-sectional pattern with 0.85 mm × 0.022 mm was achieved. In this pattern, the multilayer ceramic coil that has the internal resistance of 1 ohm was achieved. Therefore, the helical coil structure that has the magnetic core and low internal inductance was realized.

The multilayer magnetic circuit for the millimeter scale MEMS (Micro Electrical Mechanical System) air turbine generator is proposed in this paper. The dimensions of the fabricated air turbine generator were 3.6 mm, 3.4 mm and 3.5 mm, length, width and height, respectively. The air turbine was fabricated by the MEMS technology. Multilayer magnetic circuits were fabricated by the green sheet process. The achieved output voltage and output power of the generator were 6.2mV and 1.92μVA respectively. Moreover, the optimization of the ceramic magnetic circuit for the generator was performed to improve the output power. In this experiment, the horseshoe shape circuit and step-wise shape circuit were compared on the output power by the spindle machine. When two kinds magnetic circuit were compared, the output power of the step-wise shape circuit was higher than that of the horseshoe shape circuit. The output voltage and the output power of the step-wise shape circuit were 28mV and 1.53 mVA when load resistance of 0.512 Ω was connected.