In the present paper, a dynamic analysis is presented, comparing three snake-like robot gaits: lateral undulation, sidewinding locomotion, and sinus-lifting motion. To simplify calculations, locomotion motion are considered planar movements. Vertical movements assumed to be small but play critical role in change where contacts made. Thus, normal forces acting on grounded links torques applied pitch joints can calculated by solving equilibrium equations. The tradeoff between speed energy...

This brief considers the problem of trajectory tracking a planar snake robot without lateral constraint. The reference head position and orientation link 1 are given, torque control is determined to reduce errors. performance controller was tested in number simulations. robustness during actuator failure also studied. We assumed that one actuators broken corresponding joint became passive. Furthermore, as more realistic situation, we considered an instance when some states were not readily...

In many engineering problems, including those related to robotics, optimization of the control policy for multiple conflicting criteria is required. However, this can be very challenging because existence noise, which may input dependent or heteroscedastic, and restrictions regarding number evaluations owing costliness experiments in terms time and/or money. This paper presents a multiobjective algorithm expensive-to-evaluate noisy functions robotics. We present method model selection...

Abstract A research project on chemical AI, called the Molecular Cybernetics Project , was launched in Japan 2021 with goal of creating a molecular machine that can learn type conditioned reflex through process classical conditioning. In this project, we have developed learning method for network such machines, which is reported paper. First, as model machine, formulate logic gate and introduce gates. Then derive key principle learning, flipping principle, by present algorithm to realize...

The morphology and controller design of robots is often a labor-intensive task performed by experienced intuitive engineers. Automatic robot using machine learning attracting increasing attention in the hope that it will reduce workload result better-performing robots. Most are created joining several rigid parts then mounting actuators their controllers. Many studies limit possible types to finite set computational burden. However, this not only limits search space, but also prohibits use...

Abstract In this paper, we prove the existence of a reservoir that has finite-dimensional output and makes computing model universal. Reservoir is method for dynamical system approximation trains static part but fixes called reservoir. Hence, advantage training models with low computational cost. Moreover, fixed reservoirs can be implemented as physical systems. Such have attracted attention in terms computation speed energy consumption. The universality its ability to approximate an...

Central pattern generators (CPGs) play a crucial role for animal locomotion control. They can be entrained by sensory feedback to induce proper rhythmic patterns and even store the through connection weights. Inspired this biological finding, we use four adaptive neural oscillators with synaptic plasticity as CPGs control of our real snake-like robot screw-drive mechanism. Each oscillator consists only three neurons uses mechanisms based on frequency adaptation Hebbian-type learning rules....

In many robotic problems, optimization of the policy for multiple conflicting criteria is required. However this very challenging due to existence noise, which may be input dependent, or heteroscedastic, and restriction in number evaluations, experiments are expensive time and/or money. This paper presents a multiobjective (MOO) algorithm expensive-to-evaluate noisy functions robotics. We present method model selection between heteroscedastic standard homoscedastic Gaussian process...

This paper presents a kinematic calculation and control method for an inextensible continuum planar snake robot. The robot is assumed to move without side slipping; this constraint makes it easy construct its model based on which we can analyze the movement of expressed as semi-linear partial differential equation (PDE). We discuss general solution PDE it. However, also raises problem singular posture. singularities has been addressed with serial link structure; however, robot, much less...

Abstract In event‐triggered control, a situation where the control input must be sparse often arises. Therefore, in this study, we propose , meaning that is and updated an manner. First, present model‐based method for of linear systems, event condition defined by Lyapunov function. The resulting proven to system confirmed asymptotically stable. Second, extend it data‐driven version, adaptively from online data on state trajectory. Finally, discuss possibility extending our framework two...

Motor primitives provide a modular organization to complex behaviours in both vertebrates and invertebrates. Inspired by this, here we generate motor for snake-like robot with screw-drive units, thence chain combine them, order versatile, goal-directed locomotion the robot. The behavioural of are generated using reinforcement learning approach called “Policy Improvement Path Integrals” (PI2). PI2 is numerically simple has ability deal high-dimensional systems. Here, used learn robot's...

In this paper we apply a policy improvement algorithm called Policy Improvement with Path Integrals (PI <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) to generate goal-directed locomotion of complex snake-like robot screw-drive units. PI is numerically simple and has an ability deal high dimensional systems. Here, approach used find proper control parameters, like joint angles velocities, the robot. The learning process was achieved...

Path integral policy improvement (PI2) is known to be an efficient reinforcement learning algorithm, particularly, if the target system a high-dimensional dynamical system. However, PI2, and its existing extensions, have adjustable parameters, on which efficiency depends significantly. This article proposes extension of PI2 that adjusts all critical parameters automatically. Motion acquisition tasks for three different types simulated legged robots were performed test efficacy proposed...

The highly redundant nature of a snake robot will enable the to complete its task even with passive joint caused by an actuator failure. However, existing control methods cannot be applied joint. This paper proposes law that achieves linear path-following effectiveness our method is shown via numerical simulation and experiment.

Controllability Gramians are the most widely used measures of controllability, which quantify reachable state sets systems with a control input unit energy. They utilized as performance index in various design problems, such actuator placement and model reduction. If mathematical is available for system to be considered, controllability Gramian can easily calculated by its definition or equivalent conditions. However, it not always case due insufficiency data modeling. In case, practical use...

In this paper, a dynamical analysis of sidewinding locomotion by snake like robot is presented. Sidewinding treated as planar movement for simplicity, and the normal forces acting on grounded links torques applied pitch joints are calculated solving equilibrium equations. Energy efficiency compared with that lateral undulation. Simulation results show is, generally, energetically more efficient, in case where viscous coefficient small, there lower bound mean velocity center mass to be efficient than

In event-triggered control, we often face the situation where control input must be sparse. We consider sparse in which is and updated an manner. this paper, propose a data-driven method for unknown linear systems. First, as preliminary result proposed method, present constructing quadratic Lyapunov function stable system; utilizes data set of several segments state trajectories. Based on result, derive control. event-triggering condition, defined by function, online trajectory. The are...

A major part of electric power is supplied by thermal stations, which emit carbon dioxide (CO2) while generating electricity. Today, since we consume a lot electricity, are threatened environmental problems including global warming. One the key points for solving these developing energy-saving appliances. Televisions, widely used in home, can achieve energy savings reducing brightness screen. However, unless user adjusts it, TV maintains same brightness. Furthermore, tend to watch doing...

Abstract Two dynamic manipulability criteria of a snake robot with sideways slipping are proposed the application to head trajectory tracking control in mind. The singular posture, which is crucial control, characterized by and examined for families typical shapes. Differences postures from those lateral constraints, have not been clear previous studies, clarified analysis. In addition examination local properties using concept manipulability, we discuss effect isotropic friction as global...

To achieve a balanced walking for humanoid, it is necessary to estimate the dynamic stability of system. However, in small size humanoid with restricted system resource, hard satisfy performance level desired by dynamics analysis. Therefore, this paper, we propose feasible methods generate gait pattern and stabilize based on coupled oscillators which have clear correlation between oscillator parameters without real time ZMP calculation. The proposed method was tested open platform DARwIn-OP...

The ultimate goal of this research study is to perform continuous rather than sequential movements prismatic joints for effective motion a snake robot with in complex terrain. We present herein control method robotic step climbing. This composed two parts: the first involves shift reference generator that generates joint climbing step, and other use trajectory tracking controller, which head track target trajectory. In method, are divided into those directly controlled represented as...

The equipment and the structures in steel works are used a very long term. Therefore, inspections maintenance indispensable. To make this more effective, much effort has been devoted to develop methods of fault detection diagnosis. However, effectiveness those may be limited due tradeoff between false positive negative reactions. mitigate tradeoff, paper considers use model sets involving parameters disturbances that expected change. More specifically, we first propose method estimate...