Based on the analysis of the interaction between a manipulator's hand and a working object, a model representing the constrained dynamics of the robot is first discussed. The constraint forces are expressed by an algebraic function of states, input generalized forces, and constraint condition, and then direct position/force controller without force sensor is proposed based on the algebraic relation. To give the grinding system the ability to adapt to any object shape being changed by the grinding, we added estimating function of the constraint condition in real time for the adaptive position/force control, which is indispensable for our method instead of not using force sensor. Evaluations through continuous shape-grinding experiment by fitting the changing constraint surface with spline functions, indicates that reliable position/force control and shape-grinding work can be achieved by this proposed controller.
This paper proposes a new device to capture dense point cloud of human contour using horizontal LIDAR and vertical LIDAR with mirrors on pan rotational mechanism. This paper also reports a new method to estimate simple behavior such as standing and sitting online from only single vertical scan data. The combination of rotation mechanism and vertical LIDAR enlarge measurement area to overall room. Both mirrors use and pan rotation control enhance density of point cloud. The horizontal LIDAR provides robust tracking of human without missing. The experiments about performance of the device and the method demonstrated that our device and method contain sufficient feasibility to monitor human behavior in human living environments. The authors also demonstrated online recognition of simple human behavior such as standing and sitting from only single vertical scan points using pattern recognition technique.
The worldwide prevalence of obesity is considered as a serious issue because obesity is one of the main causes of diabetes, heart disease, and cancer. Obesity is associated with various habits of our daily life, e.g. short sleep duration, high alcohol intake, and nonparticipation in physical exercise. Therefore, changing these habits in order to reduce body weight is recognized as an effective solution. There are some websites in effort to monitor obesity, which verify the usefulness of web-based system for health-care support. We, the authors, propose an obesity prevention system that helps users to change their lifestyles with a website and health care devises. The system detects whether each user has the habits that generally considered as risk factors of obesity. After the detection, the system analyzes personal factors of weight gain, and informs it to the user for changing the habits effectively. Since calorie intake is an important factor of weight gain, we have developed 3 different types of dietary information logging methods. In addition, we evaluated and compared these methods by experiments, and we found that simple methods with a piece of data is enough to estimate calorie intake, instead of detailed methods with a lot of data. Moreover, we conducted another experiment for evaluating the factor analysis by predicting fluctuation of fat mass. In the experiment, the detected factors were useful for predicting the fluctuation of fat mass, therefore, we think these factors can be used for preventing obesity.
This paper presents an energy saving method for industrial machines using simple motion trajectories such as trapezoidal and S-curve velocity profiles. These velocity profiles consist of acceleration, constant velocity and deceleration time periods. This study deals with a feed drive system in which typical dynamics includes the inertia, viscous friction, Coulomb friction and back-EMF terms. The consumed energy depends only on an acceleration time period by giving total motion distance and total motion time, and by assuming that acceleration and deceleration times are identical. The authors also present an identification method of dynamic parameters of industrial machines for consumed energy prediction. Experimental results with industrial machines demonstrate the validity of the proposed analysis.
This paper provides one of the practical and meaningful applications of controller parameter tuning in the Two-Degree-of-Freedom (2DOF) control architecture for non-minimum phase systems. The paper proposes a simultaneous attainment of a desired controller and a mathematical model of a plant by utilizing fictitious reference iterative tuning (FRIT), which is useful controller parameter tuning with only one-shot experimental data, to feedforward controller in the 2DOF controllers. The proposed cost function to be minimized leads to the attainment of both of the desired controller and the mathematical model of the plant. Finally, a numerical example to show the validity of the proposed method is illustrated.
In this paper, a new simple decoupling control scheme with a model following control is proposed for 2-link industrial manipulators. Based on a coupling model, we design a compensation torque to eliminate the coupling force between two links of a manipulator in a feed-forward controller and employ observers to estimate unmodelled dynamics errors and external disturbances. Making use of the estimations, this proposes a compensation method for unmodelled errors and external disturbances to ensure the robustness of the designed control system. Effectiveness of the proposed decoupling control scheme has been confirmed through some simulation results.
Although the framework of linearly solvable Markov decision processes (LMDPs) reduces the computational complexity in reinforcement learning, it requires the knowledge of the state-transition probability in the absence of control or passive dynamics. The passive dynamics can be estimated by a temporal difference method called Z learning if the environment obeys the passive dynamics. However, it leads to a slow convergence of learning since no control is allowed during learning. This paper proposed a method to estimate the passive dynamics using Z learning under a different state-transition probability from the passive dynamics. The proposed method requires only the knowledge on what states can be visited from each possible state, and estimates the state-transition probability as well as the immediate cost of the states from the constraints they should satisfy. The computer experiments showed that the proposed method remains more efficient than Q learning with successful estimation of the passive dynamics and state costs and has a comparable convergence speed with the traditional Z learning.
In this paper, the authors presents an eco-driving nonlinear model predictive control (MPC) approach for the energy management problem of a power-split hybrid electric vehicle (HEV) system during car following. This paper adds four new contributions to this field. First, the proposed method optimizes fuel economy under the HEV physical constraints that include the upper bounds of the speed and torque of engines, motors and generators and the battery state of charge at each time. Second, in the proposed method the performance index is designed in a systematic way, which can be easily understood by designers. Third, the proposed method gives the freedom of vehicle spacing between the preceding vehicle and the host one. Fourth, using the HEV property, the desired battery state of charge is designed according to the road slopes for better recuperation of free braking energy. Computer simulation results showed that the fuel economy was much better using the nonlinear model predictive control approach than using the ADVISOR rule-based approach. The authors conclude that the nonlinear model predictive control approach is effective for the energy management problem of the power-split hybrid electric vehicle system during car following.
In this paper, the authors present a markerless vision-based estimation and control algorithm for a micro helicopter with a wireless camera. The camera looks downward, and it captures an image of the ground at each time step. The proposed algorithm consists of vision-based estimation and switched control. The 3D location and posture of the helicopter are estimated by using template matching with automatic reference generation. The proposed algorithm also detects noisy images that are sometimes captured by the camera. The control signals are produced by a set of PID controllers if the current captured image is noiseless, and otherwise by a set of integral controllers. Long-time hovering is achieved over a high contrast and planar ground. No specified markers or prior texture information are required.