Operation of model-based sensorless control of Alternating Current machines at low and zero speeds is unreliable and can fail. To overcome the limitations of sensorless control at low speeds, several alternative techniques have been developed to estimate speed and position. These are mainly based on detecting machine saliencies by measuring the response of the current to some form of voltage injection. This paper discusses injection methods, machine saliencies, and techniques used to extract speed and position that are applicable to both induction machines and permanent magnet synchronous motors.
This paper describes application trends of sensorless AC motor drives in Europe, where strict standards on energy saving are in effect. To meet their standards, each producer of AC motor drives in Europe has own strategy with specific machine design and control techniques. Generally, variable torque applications such as fans, pumps and compressors account for approximately 70% of the European market. These applications employ position and speed sensorless drives because of the advantages of reduced cost and volume, and increased reliability due to the elimination of the position sensor and related cabling connections. In this paper, first the market trends of AC motor drives in Europe are described. Second, the market penetration of sensorless AC motor drives is briefly discussed. Finally, some real-life applications are introduced.
This paper proposes a method for the fast initial position estimation of IPMSMs (interior permanent magnet synchronous motors) which is based on intentional pulse voltage injection and comb filters. Generally, the rotor position can be estimated by detecting the high-frequency current caused by the injected high-frequency voltage because its response includes inductance information due to saliency. In general, the sinusoidal current, sinusoidal voltage, and pulse voltage are utilized for saliency excitation. Methods based on the sinusoidal current and voltage, however, would require a shorter control period for precise signal injection so that methods based on pulse voltage injection would be desirable. The methods based on pulse voltage injection generally utilize the amplitude or differential value of the high-frequency current. These algorithms to calculate the saliency with its amplitude or differential value, however, appear to be complicated and unsuitable to improve the response of the position estimation. In particular, algorithms for calculating the amplitude values require some low-pass filters (LPFs) with a low cut-off frequency for the amplitude calculation, which degrades the response of the initial position estimation. In order to overcome this problem, this paper proposes a new algorithm using comb filters that can rapidly calculate the amplitude values of the high-frequency current and improve the initial position estimation performance.
This paper discusses maximum torque per ampere (MTPA) and maximum efficiency control methods based on the Volt per Hertz (V/f) control for an interior permanent magnetic synchronous motor (IPMSM). The V/f control is inherently a position sensorless method, and therefore it is simpler than conventional methods such as the sensorless vector control method. In addition, the MTPA and the maximum efficiency controls can be achieved by controlling the reactive power without requiring knowledge of the magnet pole position. The MTPA control can reduce the copper loss in IPMSMs and achieve high efficiency. Furthermore, the maximum efficiency control further improves the efficiency after implementation of the MTPA, which tends to reduce both the copper and iron losses. In this study, the validity of the MTPA methods is confirmed by simulation and experimental results. From the experimental results, the output current is reduced by 76% after the MTPA is implemented to control the reactive power. In addition at 0.6p. u. torque and 1.0p. u. motor speed, the simulation results demonstrate that the maximum efficiency control can further reduce the total losses.
The direct torque control (DTC) of ac motors leads to a faster torque response with a small number of switching operations in inverters when compared with a conventional approach such as vector control. DTC exhibits a hybrid nature in the sense that the system is composed of continuous variables of torque and flux and involves discrete switching in the inverter. The output of the inverter is limited to the finite discrete values at each instance of sampling. Model predictive control (MPC) is applied to the system so that an optimal switching sequence is derived subject to the given constraints. The proposed MPC-based approach reduces the torque ripple with a small number of switching operations. The effectiveness of the proposed MPDTC approach is verified through simulations and experiments by comparing the proposed approach with conventional DTC.
A permanent magnet synchronous motor (PMSM) rotational angle sensorless control system can have a problem with restarting when coasting. A coasting restart method has been proposed that uses the information of the three-phase short circuit current. However the coasting restart conditions of the one-time three-phase short circuit method have not been revealed. This study examined the restart conditions for a rotational angle sensorless PMSM when coasting. In the experimental tests, the duration of the three-phase short circuit was varied to examine its effect on the estimated rotor angle and rotor speed. The experimental test results were numerically analyzed to quantify the effect of the short circuit duration, because the error was caused by linearization of the estimation equation. We propose an enhanced method that considers the time response of the three-phase short circuit current. The proposed method was verified experimentally.
This paper presents a new simplified speed-sensorless vector control method for induction motors. The output voltage of the d-axis PI current controller with a decoupling control is used to compute the flux angle and to control the speed in correspondence with its reference. System stability is discussed by the root loci computed from a linear model. The effectiveness of the proposed method is demonstrated by nonlinear simulations and experiments.
In order to perform intelligent tasks using a tool, it is necessary to estimate the relation between a robot system and the tasks. However, no estimation method for this relation is currently available. Therefore, this paper proposes an estimation method for this relation by using an external sensor. In this paper, the hybrid control method based on the oblique coordinate system is utilized. Then, this relation can be treated as a task Jacobian matrix. By using the proposed method, it is possible to estimate the correct task Jacobian matrix, even if the kinematic relation between the robot system and the tool is not known or is changed during the task. As a result, the tasks are performed with this tool. From the simulation and the experimental results, the validity of the proposed method was confirmed.
Technology to record and load human motions has recently attracted attention in manufacturing. The motion-copying system saves and reproduces human motions based on position and force information. In some cases, reproducing the saved motion is difficult. This paper focused on the environmental location to propose a method of reproducing the saved motion in different locations through the effective use of motion data memory. The proposed method was verified to be able to reproduce the force and position of saved motions.
A technology that enables sharing of haptic information between several remote systems is required. A multilateral control system can implement the technology effectively. However, haptic information tends to deteriorate under communication delay because haptic information flow is bilateral. Because a conventional multilateral control connects all subsystems for information sharing, the information is greatly affected by time delay. Selecting the communication link is one way of solving the problem, however there was no means of evaluating the strength of the link. Therefore, two steps are proposed in this paper. First, an index called the “information volume” is proposed to quantify the amount of information in each subsystem. This index can be applied to a system that is experiencing time delay in its communication links. Using the index, the importance of each communication link can be quantified. Second, the algorithm to select the communication links is proposed based on the information volume. The target relationship regarding the information volumes is set, and unnecessary links are removed. The validity of the proposal is confirmed by experiments, and the results show that the degradation of haptic sensation under communication delay is prevented.
This paper presents design method for a service robot's motion and an evaluation method for impressions of these motions. The procedure consists of three phases: i) 3D CG modeling and the design of motion types, ii) subjective analysis using a two-step semantic differential (SD) method, and iii) brain monitoring to objectively evaluate the subjective answers. In the first phase, several types of robot motion are prepared by considering the psychology of Anna Freud, the Kestenberg movement profile, and personal space. In the second phase, the impression of each motion is analyzed using the SD method and factor analysis (FA). Based on the FA results, motions are classified accoring to whether they provide good or bad impressions. In the third phase, objective differences in brain activity are investigated for good and bad impressions. The result showed that the changes in cerebral blood flow at Fp1 of the brain statistically differed for good and bad impressions as classified by the subjective analysis (p < 0.01). Thus, the suitability of the presented prototyping approach to robot motion design and impression analysis was validated.
In this paper, we propose a robust vehicle tracking method based on speeded-up robust features (SURF) feature matching in a particle filter framework. In this framework, the color feature and the local binary pattern (LBP) texture feature are also combined to improve the representation of the tracking target. To further improve the tracking performance, three strategies are used. First, a dynamic update mechanism of the target template is proposed to capture appearance changes. Second, the size of the tracking window is also modified dynamically by balancing the weights of three feature distributions. Third, the weight of each particle is allocated with an improved distance kernel function method in the tracking process. Specifically, the proposed method of adopting new feature points for the target template can objectively reflect tracking target changes and effectively overcome the disadvantages of the random selection mechanism. We test the proposed approach on numerous sequences involving different types of challenges, including variations in illumination, scale changes, and rotation. The experimental results show that the proposed method is more efficient and robust than the classical approaches.