Search Results (4)

The mental fatigue of crane operators can pose a serious threat to construction safety. To enhance the safety of crane operations on construction sites, this study proposes a rotary switch semi-dry electrode for detecting the mental fatigue of crane operators. This rotary switch semi-dry electrode overcomes the problems of the large impedance value of traditional dry electrodes, the cumbersome wet electrode operation, and the uncontrollable outflow of conductive liquid from traditional semi-dry electrodes. By designing a rotary switch structure inside the electrode, it allows the electrode to be turned on and used in motion, which greatly improves the efficiency of using the conductive fluid and prolongs the electrode’s use time. A conductive sponge was used at the electrode’s contact end with the skin, improving comfort and making it suitable for long-term wear. In addition, in this study, the multifractal detrend fluctuation analysis (MF-DFA) method was used to detect the mental fatigue state of crane operators. The results indicate that the MF-DFA is more responsive to the tiredness traits of individuals than conventional fatigue detection methods. The proposed rotary switch semi-dry electrode can quickly and accurately detect the mental fatigue of crane operators, provide support for timely warning or intervention, and effectively reduce the risk of accidents at construction sites, enhancing construction safety and efficiency. Full article

Reinforcement learning has recently been studied in various fields and also used to optimally control IoT devices supporting the expansion of Internet connection beyond the usual standard devices. In this paper, we try to allow multiple reinforcement learning agents to learn optimal control policy on their own IoT devices of the same type but with slightly different dynamics. For such multiple IoT devices, there is no guarantee that an agent who interacts only with one IoT device and learns the optimal control policy will also control another IoT device well. Therefore, we may need to apply independent reinforcement learning to each IoT device individually, which requires a costly or time-consuming effort. To solve this problem, we propose a new federated reinforcement learning architecture where each agent working on its independent IoT device shares their learning experience (i.e., the gradient of loss function) with each other, and transfers a mature policy model parameters into other agents. They accelerate its learning process by using mature parameters. We incorporate the actor–critic proximal policy optimization (Actor–Critic PPO) algorithm into each agent in the proposed collaborative architecture and propose an efficient procedure for the gradient sharing and the model transfer. Using multiple rotary inverted pendulum devices interconnected via a network switch, we demonstrate that the proposed federated reinforcement learning scheme can effectively facilitate the learning process for multiple IoT devices and that the learning speed can be faster if more agents are involved. Full article

Linear flux switching permanent magnetic (LFSPM) machines, with the armature windings and magnets both on the mover in addition to a robust stator, are a good choice for long stoke applications, however, a large cogging force is also inevitable due to the double salient structure, and will worsen the system performance. Skewing methods are always employed for the rotary machines to reduce the cogging torque, and the rotor step-skewed method is a low-cost approximation of regular skewing. The step skewed method can also be applied to the linear machines, namely, the stator step skewed. In this paper, three stator step skewed structures, which are a three-step skewed stator, a two-step skewed stator and an improved two-step skewed stator, are employed for the cogging force reduction of two types of LFSPM machines. The three structures are analyzed and compared with emphasize on the influence of the skewed displacement on the cogging force and the average thrust force. Based on finite element analysis (FEA), proper skewed displacements are selected according to maximum difference between the reduction ratio of the cogging force and the decrease ratio of the average thrust force, then, the corresponding results are compared, and finally, valuable conclusions are drawn according to the comparison. The comparison presented in this paper will be useful to the cogging force reduction of LFSPM machines in general. Full article

In this paper, we will propose the neural networks integrated circuit (NNIC) which is the driving waveform generator of the 4.0, 2.7, 2.5 mm, width, length, height in size biomimetics microelectromechanical systems (MEMS) microrobot. The microrobot was made from silicon wafer fabricated by micro fabrication technology. The mechanical system of the robot was equipped with small size rotary type actuators, link mechanisms and six legs to realize the ant-like switching behavior. The NNIC generates the driving waveform using synchronization phenomena such as biological neural networks. The driving waveform can operate the actuators of the MEMS microrobot directly. Therefore, the NNIC bare chip realizes the robot control without using any software programs or A/D converters. The microrobot performed forward and backward locomotion, and also changes direction by inputting an external single trigger pulse. The locomotion speed of the microrobot was 26.4 mm/min when the step width was 0.88 mm. The power consumption of the system was 250 mWh when the room temperature was 298 K. Full article