A calculation method of polynomially parameter-dependent LMIs (Linear Matrix Inequalities) is proposed by employing Bernstein polynomial basis. Based on the fundamental properties of the Bernstein basis, it is shown that the parameter-dependent LMI conditions are systematically transformed to standard LMI conditions, and the finite division of the parameter range enables to treat the necessary and sufficient condition with arbitrary resolution and precision. The features of the proposed calculation method are illustrated with numerical examples.

Robustness of localization is a crucial issue for a mobile robot moving in various kind of environment. We have researched a robust localization system with various characteristic sensors. In past research, we proposed a selective sensor fusion system with majority rule; representing each sensor's robot position as a probability distribution, evaluating similarities between probability distributions, and selecting probability distributions based on majority rule. Although the effectiveness of the selective sensor fusion system was experimentally confirmed, proper selection could not be done in some case because similarity evaluation had a problem. This paper proposes up-to-date system to conduct more proper sensor selection. Simulational and experimental results show that the proposed system can select proper sensors in various environment.

Spacecraft landing gear employed in space missions is required to achieve secure touchdown on rough and inclined terrains. Generally, when a spacecraft performs a free fall from a certain altitude, its landing gear needs to absorb the impact of touchdown. Conventional landing gear, such as the honeycomb crush absorber, absorbs the impact of landing through plastic deformation of its structure. However, such landing gear cannot effectively prevent the spacecraft from tipping over, and the non-reusability of such landing gear often leads to an increase in experimental costs. To address these issues, this paper proposes a novel landing gear mechanism that comprises a contraction lock mechanism with multiple springs for enhancing reusability. The proposed mechanism varies the spring constant by operating the contraction lock mechanism according to the touchdown response, and thus potentially prevents the spacecraft from tipping over. The effectiveness of the proposed mechanism in the case of inclined terrains is verified through conducted simulations. Furthermore, the performance of the proposed mechanism is compared with that of the conventional plastic deformation shock absorber in terms of adaptability to variations in the spacecraft's initial velocity and initial attitude angle. The obtained results show that the proposed mechanism can be effective in executing secure landings on inclined terrains.

Many experimental devices employing DC motors have been used to study control engineering. The model of them can be approximated as a linear system and we can control the angle or the angular velocity of a DC motor sufficiently by using classical control theory, like PID control. However, just rotating a motor does not seem to attract the attention of students in mechanical engineering classes. Thus, an experimental curriculum employing a mobile robot is proposed to learn control theory. The kinematic model of the mobile robot includes non-linearity. In this paper, two representations are used to analyze and design a controller for the nonlinear mobile robot: a linear approximated model and a linear model obtained by a strict linearization method. The responses when a PID controller is applied are analyzed based on the linear approximated model, and the limits for controlling the mobile robot using a PID controller are shown. When the nonlinear controller designed based on a strict linearization method is used, the control gains can be adjusted based on linear control theory, that is, the pole assignment method. Students can recognize the relationship between the poles and responses of the controlled system as the traveling path of a mobile robot. The proposed experiments were actually performed in some classes and attracted the attention of the students.

This paper proposes a cooperative formation control of multi-agent systems with self-equilibrium force by using parallel virtual springs connecting the agents. The method is an extension of our previous work focused on the formation control with the self-equilibrium force, where the control force is generated from a single virtual spring for each pair of the agents. The previous control guarantees the local stability, but not the global stability. Hence it will be problematic when agents are interspersed away from the target position. This paper is devoted to solve this problem by introducing multiple virtual springs between the each pair. In addition to the formation and self-equilibrium force control, collision avoidance can be also considered by introducing an additional virtual spring. Effectiveness of the method is evaluated by numerical examples.

For accurate global self-localization, researches for the feature description of the laser-scan data have been actively conducted. Main approaches to the feature description are to design feature descriptor based on human knowledge regarding the specific environment, e.g., office and hallway. However, in real robot navigation tasks such as a security patrol robot, the robot would be applied to a variety of environments and it is expensive if the users need to tune the design at every environment. To alleviate such problem, we propose to extend the state-of-the-art variational auto-encoder (VAE) by introducing the step-edge detector, which detects non-continuous transition emerged frequently at the laser scan data due to the limitation of distance measurement. With our proposed method, called “LaserVAE”, the feature descriptor of the laser scan is automatically tuned given unknown environments. Through experiments with a real self-localization with a 2D laser scanner, we demonstrate the effectiveness of the proposed method.