Artificial Life and Robotics最新文献

With the aim of allowing the efficient and realistic simulation of swarm algorithms for exploration and coverage, we present the tool Multi-Agent Exploration Simulator (MAES), which is an open-source physics-based discrete step multi-robot simulator. MAES features movement in a continuous 2D space, realistic physics based on the Unity framework, advanced visualization techniques such as heatmaps, custom wireless signal degradation, both randomly generated and custom user-provided maps, and a ROS (Robot Operating System) interface. This latter characteristic could allow to port the simulated algorithms to real-world robots. We present performance tests, conducted with rather modest hardware, showing that MAES is able to simulate up to 5 robots in ROSMode (using the ROS integration) and up to 120 robots in UnityMode (development performed directly into the C# Unity Editor). A usability test was conducted which hinted that the target audience of robotics researchers and developers is able to quickly install, setup, and use MAES for implementing simple robot logic.

Based on the human–robot interaction behavior of mobile robots in social navigation, this paper proposes a social navigation quality evaluation model based on combined weights for the problems of single indicators, rough quantification and non-convergence of information in social navigation quality evaluation. Firstly, three evaluation modules of comfort, naturalness and sociality are designed, and each module is refined into primary and secondary indicators. The robot path navigation data are calculated by the indicator quantification formula. Secondly, the subjective and objective weights of hierarchical analysis method and the entropy weight method are combined to determine the index weights at each level. The weighted sum is used to achieve the fusion of index information and obtain the optimal solution of the evaluation navigation algorithm. Finally, we simulate the social scene through visualization simulation experiments to obtain the trajectory data of the robot in the social scene. The experimental results verify the feasibility of the theoretical model and give the final scores and optimization opinions of the tested algorithms. Through the evaluation of the social navigation quality evaluation model, the path planning algorithm that best suits the comfort perception of pedestrians in the current scenario can be found in the tested algorithms.

The problem of brain–computer interface (BCI) using steady-state visual evoked potential (SSVEP) is a flickering sensation caused by the flashing stimuli used to induce SSVEP. To use of high-frequency flashing stimuli is one of the countermeasures of this problem. This study focused on the relationship between the magnitude of SSVEP components for each subject and proposed a high-frequency (56–70 Hz) SSVEP–BCI that uses only the frequencies at which SSVEP induction was confirmed. For comparison, the accuracy of SSVEP–BCI using learning CCA (LCCA), an extension of canonical correlation analysis (CCA), was 98.61% for the low-frequency (26–40 Hz) SSVEP–BCI for comparison, 62.78% for the high frequency (56–70 Hz) SSVEP–BCI, and 87.19% for the high frequency (56–70 Hz) SSVEP–BCI with personalized stimulus frequency. As a result of comparing with and without personalization using information transfer rate (ITR), non-personalized (normal) and personalized high-frequency SSVEP–BCI ITR were 24.25 bits/min and 29.64 bits/min.

In conventional smart phones and ATMs, a four-digit passcode is entered into a keypad, and the user confirms whether the passcode matches the keypad. However, there is a risk that a third party can easily steal the password by watching the code entry or analyzing the position of fingerprints left on the keypad. There are other solutions, such as biometric authentication or the use of special displays, but both of them are costly and difficult to implement. In this study, we propose a keypad that does not leave fingerprints on the screen, is low cost, and can be used to input passcodes without worry, even if someone is standing next to it. The proposed keypad uses cursors that are moved by directional keys to select numbers, making fingerprint analysis difficult. Because attackers do not know the color that the user has selected, they cannot know which cursor the user is moving. To verify the safety and convenience of this system, we conducted experiments on subjects in their 20 s and 50 s. The results showed that the average difference in authentication time from the conventional method was about 5 s, and the method was generally convenient. We conclude that our keypad system is secure, because no peeping attacks on a subject were successful in guessing the subject’s passcode.

One of the new computational frameworks is physical reservoir computing. Focusing on this method, we have previously developed a soft-matter artificial mouth ”Gel Biter”, which is composed of multiple polymeric materials based on the structure of the human oral cavity. This soft machine can discriminate even subtle differences in food texture with high accuracy. In general, chewing speed differs from person to person. Then, we focus on the result that brittle foods tend to be chewed faster or more finely based on sensory evaluation in some cognitive studies. This study has analyzed the accuracy of the Gel Biter by changing the parameters of its robotic arm and the differences in food texture perceived when the chewing speed is changed. As a result, there is no significant difference in discrimination accuracy for each speed. The cluster analysis shows that the food characteristics are captured and classified. In addition, the estimation results for Fast chewing indicate that the mechanical mouth also generates the illusion that humans perceive different food textures.

Mushrooms grow so fast during the harvest season that they can double in size in a day. However, the soft and fragile nature of mushrooms makes manual harvesting of domestic brand varieties a necessity. Therefore, an important industrial issue in the efficiency of mushroom cultivation in Japan is how to make mushrooms easy to harvest and grow. The technical elements of mushroom harvesting are (1) non-damaging harvesting methods, (2) control of colony growth, and (3) expansion of growing area. This study proposes a three-dimensional and deformable culture medium to solve the problems (1)–(3). The proposed three-dimensional medium has a 3D-printed anisotropic elastic well structure embedded inside. The medium keeps the medium in a three-dimensional shape and allows mushrooms to be generated from the sides and bottom. In addition, we show that during the harvesting period, by applying pressure to the medium in a single direction. The soil can be removed from each side of the mushrooms and the mushrooms can be harvested.

Novel basic analog–digital edge detection circuits in this research were proposed based on the vertebrate retina. The vertebrate retina is a pre-processor for image processing in the brain and has superior functions such as edge detection and motion detection. The proposed circuit was evaluated by the simulation program with integrated circuit emphasis (SPICE) with the 0.6 μm complementary metal oxide semiconductor (CMOS) process. The simulation results with SPICE showed that the proposed circuits can operate normally. The test integrated circuit of the unit circuit was fabricated with a 0.6 μm CMOS process. The measured results showed that the unit circuit can detect the edge position. The test integrated circuits of one-dimensional array of unit circuits were fabricated with the same process. The measured results showed that the one-dimensional circuits can detect the edge position. In the future, the realization of a new vision sensor can be expected using the proposed circuit.

Manipulation for deformable object is difficult in robotics. The deformation of the deformable object is not the same, despite the same manipulation. This is due to the difference in the object characteristics, which depend on knitting, material, etc. This leads to difficulties in the motion planning. We propose a method that estimates the string model by comparing the real string movement and simulated string movement in a certain manipulation repeatedly by trial and error. This method realizes several manipulations using unknown strings. But feasible range was limited to uniform strings. In this paper, we proposed string model for representing various kind of string. This model assumed that mass distribution is not uniform and bending properties is different depending on extraction and contraction. Where this model was applied to several non-uniform string and uniform string, we confirmed that the proposed model can express the actual string movement.

Japan is one of the countries in the world where natural disasters occur most frequently. Typhoons, earthquakes, tsunamis, volcanic eruptions, and many other natural disasters occur in Japan every year, and the risks are immeasurable. Japan is especially prone to earthquakes because of its location on the North American Plate, Eurasian Plate, Pacific Plate, and Philippine Sea Plate. Preparation for a Nankai Trough earthquake is one of the most important issues. A Nankai trough earthquake is predicted to occur with a probability of 70–80% within 30 years. Accordingly, a tsunami of more than 10 ms is expected to hit a wide area along the Pacific coast from the Kanto region to the Kyushu region, and the impact is expected to be significant. When a tsunami warning is issued, an evacuation support system that provides information on evacuation sites and routes enables faster and safer evacuation actions. In a previous study conducted in our laboratory, we proposed an evacuation support system using Low-Power Wide-Area communications and conducted a simulation to find ways to improve the evacuation support system. However, the simulation was insufficient in that it did not consider delays in evacuation completion time due to road congestion caused by limited road width, which can be expected in an actual disaster. In our current research, as an improvement to make the simulation more realistic for further developing our evacuation support system, we propose a simulation that considers road width. We examined how adoption of this simulation methodology would affect the simulation results by comparing simulations that did and did not consider road width. As a result, the proposed method showed a lower percentage of completed evacuations than the previous simulation, especially for children and the elderly. The reason for this result is that when evacuations begin simultaneously, crowding occurs as people all rush to evacuation centers. This congestion on the roads near evacuation centers prevents people from completing their evacuations. Such a phenomenon is expected to occur not only in simulations but also on actual roads. For this reason, our proposed simulation method that considers road width is a more realistic simulation. Since evacuations by car in the event of a disaster also occur, a future issue is to simulate car congestion as well. Furthermore, simulating what would happen if a disaster made a road impassable is also important.

In designing and manufacturing processes using CAD/CAM systems, cutter location source (CLS) data are generally used for intermediate data to finally generate numerical control (NC) data for various types of NC machine tools. CLS data mainly include statements as ‘GOTO’ to designate the position and orientation of a cutting tool. However, unfortunately, industrial robots and mechatronics systems have not been supported by such standardized CLS data. Also, it has not been realized for CLS data to be able to optionally have special statements for such as handling a customized end-effector and a camera system, executing a visual feedback control or cooperative control, and implementing AI systems like convolutional neural networks (CNNs). This paper aims to propose hyper-CLS data named HCLS data that can support the above functions. The effectiveness of HCLS data is experimentally evaluated by implementing a sequence work process using two small-sized robots with four-DOFs. The cost required for introducing multiple industrial robots is a serious issue for small and medium-sized manufacturers. Recently, there has been a growing need for small-sized industrial robots, which are comparatively easy to be introduced at low cost, in order to automate work processes that have previously relied on human hands, such as parts fitting tasks with a small clearance and cooperative task taking the timing. The proposed cooperative system based on two small-sized four-DOFs articulated robots enables a sequence control for picking and placing without using any programmable logic controller (PLC). The effectiveness and the validity of the proposed robot system are demonstrated through a pick and place experiment.