• The Journal of Korea Robotics Society
• /
• v.19 no.3
• /
• pp.266-273
• /
• 2024

With the rising interaction between robots and humans, detecting collisions has become increasingly vital for ensuring safety. In this paper, we propose a novel approach for detecting collisions without using force torque sensors or tactile sensors, utilizing a Transformer-based neural network architecture. The proposed collision detection approach comprises a torque estimator network that predicts the joint torque in a free-motion state using Synchronous time-step encoding, and a collision discriminator network that predicts collisions by leveraging the difference between estimated and actual torques. The collision discriminator finally creates a binary tensor that predicts collisions frame by frame. In simulations, the proposed network exhibited enhanced collision detection performance relative to the other kinds of networks both in terms of prediction speed and accuracy. This underscores the benefits of using Transformer networks for collision detection tasks, where rapid decision-making is essential.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.21 no.5
• /
• pp.32-43
• /
• 2007

Interior permanent magnet synchronous motor(IPMSM) has become a popular choice in electric vehicle applications, due to their excellent power to weight ratio. This paper proposes maximum torque control of IPMSM drive using adaptive learning fuzzy neural network and artificial neural network. This control method is applicable over the entire speed range which considered the limits of the inverter's current and voltage rated value. This paper proposes speed control of IPMSM using adaptive learning fuzzy neural network and estimation of speed using artificial neural network. The back propagation neural network technique is used to provide a real time adaptive estimation of the motor speed. The proposed control algorithm is applied to IPMSM drive system controlled adaptive learning fuzzy neural network and artificial neural network, the operating characteristics controlled by maximum torque control are examined in detail. Also, this paper proposes the analysis results to verify the effectiveness of the adaptive learning fuzzy neural network and artificial neural network.

• Proceedings of the Korean Fiber Society Conference
• /
• 1991.06b
• /
• pp.76-76
• /
• 1991

• KEPCO Journal on Electric Power and Energy
• /
• v.5 no.3
• /
• pp.229-233
• /
• 2019

Large-scale High Temperature Superconducting (HTS) wind power generators suffer from high electromagnetic force and high torque due to their high current density and low rotational speed. Therefore, the torque and Lorentz force of HTS wind power generators should be carefully investigated. In this paper, we proposed a Performance Evaluation System (PES) to physically test the structural stability of HTS coils with high torque before fabricating the generator. The PES is composed of the part of a pole-pair of the HTS generator for estimating the characteristic of the HTS coil. The 10 MW HTS generator and PES were analyzed using a 3D finite element method software. The performance of the HTS coil was evaluated by comparing the magnetic field distributions, the output power, and torque values of the 10 MW HTS generator and the PES. The magnetic flux densities, output power, and torque values of the HTS coils in the PES were the same as a pole-pair of the 10 MW HTS generator. Therefore, the PES-based evaluation method proposed in this paper can be used to estimate the critical characteristics of the HTS generator under high magnetic field and high torque before manufacturing the HTS wind turbines. These results will be used effectively to research and manufacture large-scale HTS wind turbine generators.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.4
• /
• pp.517-522
• /
• 2019

This paper proposes a new motor parameter estimation method. Because the proposed method is based on difference equations, it does not affect the error in the voltage magnitude so called dead-time effect. Information on the motor constant may be needed to improve the motor control performance. For example, a control technique called DTC (Direct Torque Control) requires a motor constant when calculating the torque and flux magnitude. As another example, in the case of predictive control, information on the motor parameters is required to generate voltage references. Because the constant of the motor fluctuates according to the driving environment, it is essential to estimate the correct motor constant because the control performance is degraded when incorrect motor information is used. In the proposed scheme, the motor constant estimated based on the voltage difference equation is obtained using the RLS (Recursive Least Square) technique. The RLS algorithm is applied to obtain the value through an iterative calculation so that the estimation performance is robust to noise. The simulation results carried out with surface mounted permanent magnet motors confirmed the validity of the proposed method.

• Journal of rehabilitation welfare engineering & assistive technology
• /
• v.7 no.2
• /
• pp.7-12
• /
• 2013

In this paper, we propose a brushless DC motor drive system for active knee prosthesis and low-cost estimation method for phase current from DC-link current. To control motor torque directly, current sensing is very important and current sensing point should be synchronized with voltage switching command to minimize the effect of switching noise in current measurement, For maintaining small form factor, simplifying control schemes and achieving low-cost system, control schemes using DC-link current are used. Moreover, we incorporated phase current estimation method using analog MUX for minimizing current estimation error between DC-link current and phase current. The validity of the proposed system is verified through experimental works.

• Journal of Biomedical Engineering Research
• /
• v.22 no.3
• /
• pp.293-301
• /
• 2001

A patient-specific musculoskeletal model, whose parameters can be identified noninvasively, was developed for the automatic generation of patient-specific stimulation pattern in FES. The musculotendon system was modeled as a torque-generator and all the passive systems of the musculotendon working at the same joint were included in the skeletal model. Through this, it became possible that the whole model to be identified by using the experimental joint torque or the joint angle trajectories. The model parameters were grouped as recruitment of muscle fibers, passive skeletal system, static and dynamic musculotendon systems, which were identified later in sequence. The parameters in each group were successfully estimated and the maximum normalized RMS errors in all the estimation process was 8%. The model predictions with estimated parameter values were in a good agreement with the experimental results for the sinusoidal, triangular and sawlike stimulation, where the normalized RMS error was less than 17%, Above results show that the suggested musculoskeletal model and its parameter estimation method is reliable.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2002.11b
• /
• pp.122-127
• /
• 2002

A frequency-domain PES estimation and its prediction method are proposed for the tightly-coupled servo/mechanical design of high-TPI HDD system above 100 kTPI. The major two disturbance energies which are related with mechanical vibrations inside of HDD are used to predict the drive-level PES, while considering closed-loop servo dynamics. One is the torque disturbance which mainly comes from aerodynamic excitation of HSA system and the other is the displacement disturbance from disk-spindle dynamics. In order to obtain the accurate error transfer function of closed-loop servo control, the plant model is measured by accurate experiment. The measured PES is compared with predicted one in terms of frequency-domain PES spectrum and its standard variation value. It is proved that the proposed frequency-domain PES estimation/prediction method is capable of predicting drive-level PES of high-TPI hard disk drive.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.55 no.3
• /
• pp.133-140
• /
• 2006

This paper is proposed an efficiency optimization control algorithm for a synchronous reluctance motor(SynRM) which minimizes the copper and iron losses. Also, this paper presents a speed estimated control scheme of SynRM using artificial neural network(ANN). There exists a variety of combinations of d and q-axis current which provide a specific motor torque. The objective of the efficiency optimization controller is to seek a combination of d and q-axis current components, which provides minimum losses at a certain operating point in steady state. It is shown that the current components which directly govern the torque production have been very well regulated by the efficiency optimization control scheme. The proposed algorithm allows the electromagnetic losses in variable speed and torque drives to be reduced while keeping good torque control dynamics. The control performance of ANN is evaluated by analysis for various operating conditions. Analysis results are presented to show the validity of the proposed algorithm.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.27 no.8
• /
• pp.1251-1258
• /
• 2003

This paper presents a nonlinear robust approach to the slip control problem for a torque converter bypass clutch in a passenger car. The proposed nonlinear robust controller builds upon only the measurements avail-able from inexpensive sensors that are already installed in passenger cars for control. The issue of torque estimation problems for the implementation of the proposed controller is addressed. The stability of the internal dynamics is investigated, upon which a nonlinear robust controller is designed using input-output feedback linearization and Lyapunov redesign technique. The performance of the designed controller is validated by simulation studies.