Locomotion Experiments Research Articles

A model-free method to learn multiple skills in parallel on modular robots

Legged robots are well-suited for deployment in unstructured environments but require a unique control scheme specific for their design. As controllers optimised in simulation do not transfer well to the real world (the infamous sim-to-real gap), methods enabling quick learning in the real world, without any assumptions on the specific robot model and its dynamics, are necessary. In this paper, we present a generic method based on Central Pattern Generators, that enables the acquisition of basic locomotion skills in parallel, through very few trials. The novelty of our approach, underpinned by a mathematical analysis of the controller model, is to search for good initial states, instead of optimising connection weights. Empirical validation in six different robot morphologies demonstrates that our method enables robots to learn primary locomotion skills in less than 15 minutes in the real world. In the end, we showcase our skills in a targeted locomotion experiment.

Comparative study of soil interaction and driving characteristics of different agricultural and space robots in an agricultural environment

AbstractThis paper investigates four different mobile robots with respect to their driving characteristics and soil preservation properties in an agricultural environment. Thereby, robots of classical design from agriculture as well as systems from space robotics with advanced locomotion concepts are considered to determine the individual advantages of each rover concept with respect to the application domain. Locomotion experiments were conducted to analyze the general driving behavior, tensile force, and obstacle‐surmounting capability and ground interaction of each robot. Various soil conditions typical for the area of application are taken into account, which are varied in terms of moisture and density. The presented work covers the specification of the conducted experiments, documentation of the implementation as well as analysis and evaluation of the collected data. In the evaluation, particular attention is paid to the change in driving characteristics under different soil conditions, as well as to the soil stress caused by driving, since soil quality is of critical importance for agricultural applications. The analysis shows that the advanced locomotion concepts, as used in space robotics, also have positive implications for certain requirements in agricultural applications, such as maneuverability in wet conditions and soil conservation. The results show potential for design innovations in agricultural robotics that can be used, to open up new fields of application for instance in the context of precision farming.

Continual Reinforcement Learning for Quadruped Robot Locomotion.

The ability to learn continuously is crucial for a robot to achieve a high level of intelligence and autonomy. In this paper, we consider continual reinforcement learning (RL) for quadruped robots, which includes the ability to continuously learn sub-sequential tasks (plasticity) and maintain performance on previous tasks (stability). The policy obtained by the proposed method enables robots to learn multiple tasks sequentially, while overcoming both catastrophic forgetting and loss of plasticity. At the same time, it achieves the above goals with as little modification to the original RL learning process as possible. The proposed method uses the Piggyback algorithm to select protected parameters for each task, and reinitializes the unused parameters to increase plasticity. Meanwhile, we encourage the policy network exploring by encouraging the entropy of the soft network of the policy network. Our experiments show that traditional continual learning algorithms cannot perform well on robot locomotion problems, and our algorithm is more stable and less disruptive to the RL training progress. Several robot locomotion experiments validate the effectiveness of our method.

Enhancement of humanoid robot locomotion on slippery floors using an adaptive controller

AbstractThis paper presents a comprehensive strategy to improve the locomotion performance of humanoid robots on various slippery floors. The strategy involves the implementation and adaptation of a divergent component of motion (DCM) based control architecture for the humanoid NAO, and the introduction of an embedded yaw controller (EYC), which is based on a proportional-integral-derivative (PID) control algorithm. The EYC is designed not only to address the slip behavior of the robot on low-friction floors but also to tackle the issue of non-straight walking patterns that we observed in this humanoid, even on non-slippery floors. To fine-tune the PID gains for the EYC, a systematic trial-and-error approach is employed. We iteratively adjusted the P (Proportional), I (Integral), and D (Derivative) parameters while keeping the others fixed. This process allowed us to optimize the PID controller’s response to different walking conditions and floor types. A series of locomotion experiments are conducted in a simulated environment, where the humanoid step frequency and PID gains are varied for each type of floor. The effectiveness of the strategy is evaluated using metrics such as robot stability, energy consumption, and task duration. The results of the study demonstrate that the proposed approach significantly improves humanoid locomotion on different slippery floors, by enhancing stability and reducing energy consumption. The study has practical implications for designing more versatile and effective solutions for humanoid locomotion on challenging surfaces and highlights the adaptability of the existing controller for different humanoid robots.

Design of an Amphibious Spherical Robot Driven by Twin Eccentric Pendulums With Flywheel-Based Inertial Stabilization

This article presents the design of an amphibious spherical robot with a sole focus on its mechanical design to achieve high torque, versatile motion modes, compact and lightweight structure, and high stability and traction, without relying on optimal control strategies. To achieve this, we unified the use of twin eccentric pendulums for high torque, flywheels for inertial stabilization and enhanced traction, and flywheel brakes for instantaneous torque to overcome obstacles. Dynamic analysis is performed, which gives an insight into the robot's motion and helps in making design decisions. High-level and low-level design decisions are discussed based on robot dynamics and intuition. Finally, a prototype of the designed spherical robot is assembled, and subjected to performance evaluation through terrestrial and aquatic locomotion experiments. The experimental results validated our design decisions, showing the design approach yields a spherical robot with stable and versatile locomotion, building the foundation for further improvements via optimal control and sensing.

Hydraulic variable stiffness mechanism for swimming locomotion optimization of soft robotic fish

In this paper, we describe the development of a soft robotic fish with Hydraulic Variable Stiffness (HVS), a simple and light mechanism that actively controls the tail stiffness. Modal analysis model using PRBM was developed to predict the tendency of optimal tail stiffness which maximizes the forward swimming speed in response to driving frequency. The forward swimming locomotion experiment demonstrated that the optimal tail stiffness which maximizes swimming speed exists for each driving frequency, and that optimal stiffness increases as the frequency increases. The turning locomotion experiment showed that, as the tail stiffness decreases, the turning radius decreases and the turning angular velocity increases significantly. By controlling the tail stiffness to its optimal for specific locomotion, the robot can increase its swimming speed by 118% on average, reduce its turning radius to almost a quarter, and increase its turning angular velocity by almost three times. These results demonstrate that the HVS soft robotic fish can optimize its swimming locomotion more effectively compared to conventional robots by optimizing its tail stiffness depending on its required swimming situation with only one actuator to oscillate the tail using the HVS device.

ROBOMINER: Development of a Highly Configurable and Modular Scaled-Down Prototype of a Mining Robot

Historically, mining operations have faced numerous challenges, including safety hazards, inefficiencies, and environmental concerns. However, recent advances in robotics, automation, and artificial intelligence have presented opportunities for the mining industry. The ROBOMINERS project, a Horizon 2020 European Union initiative, aims to revolutionize the mining ecosystem by implementing disruptive robotic concepts. One such concept is resilience, which involves enabling mining robots to reconfigure morphologically during operation. This article presents the development of a modular robotic system that focuses on modularity and self-assembly to provide insight into developing a highly adaptable and compact solution for future mining robots. The robotic system is composed of a set of highly configurable modular robotic platforms that can be reconfigured with other robotic modules or submodules to form more complex systems to perform different tasks. Several module configurations are presented, and different locomotion experiments were carried out to test the ability of the modules to navigate unstructured environments. The modules exhibited great maneuverability in unstructured terrain and demonstrated self-assembly and reconfiguration capabilities during operation. This is a foundational step towards the long-term goal of developing compact autonomous agents capable of self-assembly and mining task execution.

Traversability analysis for off-road environments using locomotion experiments and earth observation data

In recent years, the navigation capabilities of mobile robots in off-road environments have increased significantly, opening up new potential applications in a variety of settings. By accurately identifying different types of terrain in unstructured environments, safe automated navigation can be supported. However, to enable safe path planning and execution, the traversability costs of the terrain types need to be accurately estimated. Such estimations are often performed manually by experts who possess information about the environment and are familiar with the capabilities of the robotic system or using simplified experiments. In this paper, we present an automated pipeline for generating traversability costs that use recorded locomotion data from a realistic experiment and descriptive information on the terrain obtained from earth observation data. The main contribution is that the cost estimation for different terrain types is based on locomotion data obtained in realistic standardized experiments. Moreover, by repeating the experiments with different robot systems we are easily able to reflect the actual capabilities of the systems. Experiments were conducted in an alpine off-road environment to record locomotion data of four different robot systems and to investigate the performance and validity of the proposed pipeline. The recorded locomotion data for the different robots are publicly available at https://robonav.ist.tugraz.at/data/

Human Joint Axis Estimation Algorithm Using a Single Sensor to Recognize the Intention of the Wearer of a Wearable Robot

Wearable robots have been developed to aid or substitute human gait locomotion. To assist gait locomotion based on the intention of the wearer, gait pattern analysis is required with a wearable sensor that measures body information, such as the joint angular velocity. However, it is difficult to measure the joint angular velocity precisely because of the attachment position of a sensor that has a curvature and an anatomical human joint axis that is invisible. Therefore, a sensor calibration algorithm that aligns the sensor axis with an anatomical human joint axis is required to provide appropriate assistance to the wearer. Hence, in this study, a new and simple sensor calibration algorithm that uses a single sensor to simplify the robot system and reduce its weight is proposed. The attachment position of the sensor may change because the wearer shakes the body or collides with the ground when walking. Thus, a continuous sensor-compensation algorithm is proposed to correct for the changes in the location of the sensor. The effectiveness of the proposed algorithm is demonstrated through gait locomotion experiments on various paths. In addition to the gait locomotion experiments, the proposed algorithm is applied to a real wearable robot system. An ankle-assist wearable robot experiment reveals that the proposed algorithm can enhance the sensing and recognition performance of the robot system.

A worm-snake-inspired metameric robot for multi-modal locomotion: Design, modeling, and unified gait control

Worm-like and snake-like robots have attracted a great deal of research attention due to their slender bodies and excellent mobility. Based on different locomotion mechanisms, worm-like robots are more suited for movement in restricted environments, while snake-like robots excel in fast movements in open spaces but have limited mobility in confined areas. To complement their advantages, this paper presents a new design and prototype of a Worm-Snake-Inspired Metameric (WSIM) robot with multi-modal locomotion capability, which can execute peristaltic planar locomotion by exploiting the contracting of the worm-like modules and serpentine planar locomotion via active swing of the snake-like joints. We also propose a unified gait control framework that unifies the gait signals of worm-like and snake-like locomotion modes by using the parameter vector, and the conditions that need to be satisfied by the control parameters for generating different gaits are derived. This study also takes a major step forward in establishing dynamic models for the worm-like and snake-like locomotion modes, which can effectively predict the robot's locomotion performance, including the average velocity, slope, and radius of the trajectory. Gait experiments and complex scenario locomotion experiments demonstrate that the WSIM robot indeed has multi-modal locomotion capability and is well suited to execute tasks in complex environments including tubes, open areas, and narrow corridors. The findings of this paper would provide a useful basis for the design, modeling, and control of future bioinspired multi-modal robots.

Muscle prestimulation tunes velocity preflex in simulated perturbed hopping

Muscle fibres possess unique visco-elastic properties, which generate a stabilising zero-delay response to unexpected perturbations. This instantaneous response—termed “preflex”—mitigates neuro-transmission delays, which are hazardous during fast locomotion due to the short stance duration. While the elastic contribution to preflexes has been studied extensively, the function of fibre viscosity due to the force–velocity relation remains unknown. In this study, we present a novel approach to isolate and quantify the preflex force produced by the force–velocity relation in musculo-skeletal computer simulations. We used our approach to analyse the muscle response to ground-level perturbations in simulated vertical hopping. Our analysis focused on the preflex-phase—the first 30 ms after impact—where neuronal delays render a controlled response impossible. We found that muscle force at impact and dissipated energy increase with perturbation height, helping reject the perturbations. However, the muscle fibres reject only 15% of step-down perturbation energy with constant stimulation. An open-loop rising stimulation, observed in locomotion experiments, amplified the regulatory effects of the muscle fibre’s force–velocity relation, resulting in 68% perturbation energy rejection. We conclude that open-loop neuronal tuning of muscle activity around impact allows for adequate feed-forward tuning of muscle fibre viscous capacity, facilitating energy adjustment to unexpected ground-level perturbations.

Stable Flexible-Joint Floating-Base Robot Balancing and Locomotion via Variable Impedance Control

This article presents a framework for systematic, stable and passive, dynamical balancing and locomotion control of flexible-joint bipedal robots. In order to achieve stability/passivity of the full flexible-joint, floating-base model with contacts, several novel control designs are proposed, whose ability to guarantee regulation and tracking stability is mathematically and practically demonstrated. The proposed designs enable usage of full-state feedback terms, thereby increasing both link tracking and oscillation suppression performance. These constitute the only control schemes reported in the literature, which are capable of asymptotically stabilizing flexible-joint, floating-base systems with contacts, during trajectory-tracking tasks. Moreover, a novel linear quadratic regulator (LQR) tuning approach is proposed, which permits the creation of models characterized by distinct kinetic chain and impedance combinations. Stable switching between these gain sets is guaranteed, as it is demonstrated that the proposed controllers enable unconstrained and stable, variable impedance control (VIC). The proposed control methods are corroborated through practical, balancing and locomotion experiments on the COmpliant huMANoid, as well as via dynamical simulations; these results demonstrate stability maintenance during tracking and VIC tasks. The ability to stably modulate a legged robot's active impedances could enable closer replication of biologically inspired behaviors.

보행 보조 로봇의 보행 시점 판별 성능 향상을 위한 자이로 센서 신호 처리 기법

Wearable robots have been developed to aid or substitute the gait locomotion of humans. To assist gait locomotion based on the intention of a wearer, a gait pattern analysis is required with a wearable sensor by measuring body information, i.e., a joint angular velocity. However, measuring a precise joint angular velocity is difficult because the attachment position of a sensor has a curvature and an anatomical joint axis which is invisible. Therefore, a sensor calibration algorithm, which aligns a sensor axis into an anatomical joint axis, is required to provide an apposite assist for a wearer. Hence, in this paper, a new and simple sensor calibration algorithm is proposed with a one sensor which contributes to simplifying and lightening the robot system. Since a wearer shakes the body or collides with the ground when walking, the attachment position of a sensor may be changed. Thus, a continuous sensor compensation algorithm is also proposed. The effectiveness of this new algorithm is demonstrated by gait locomotion experiments on various paths.

Reaction Force Analysis for Obstacle-Aided Locomotion of Snake Robot Using Piecewise Helixes

Snake robots can use numerous joints to move in various manners. One of them is the obstacle-aided locomotion, a method in which a robot moves by pushing obstacles. In our previous study, we proposed a three-dimensional obstacle-aided locomotion using piecewise helixes, which is composed of an obstacle-contacting part and a grounding part. However, this method is imperfect. This is because when depending on the contact condition with obstacles, the reaction force may make it difficult for the robot to maintain contact with the obstacles or to propel itself. In this study, we analyze the reaction force received from obstacles during the obstacle-aided locomotion to improve the performance of the method. Using these analysis results, we evaluate the contact points of the obstacle-contacting part from two perspectives. The first considers the maintenance of contact with the obstacle, and the second considers the factors that interfere with propulsion. The reaction force was confirmed experimentally using a force sensor. Additionally, the theory of the factors that interfere with the propulsion and maintenance of contact with the obstacle was verified using experiments of the obstacle-aided locomotion in several environments.

A Multigait Continuous Flexible Snake Robot for Locomotion in Complex Terrain

This article presents a novel multigait continuous flexible snake mobile robot based on rotating, twisting, and untwisting of a flexible spring. The proposed robot combines the advantages of a twisted flexible helical spring and a string of rigid planar articulated circular shells into a unique wheel-less structure with 100 degrees of freedom per meter length. By adjusting the rotation direction and speed of two motors set on both ends of the constrained spring, the robot performs different motions. A kinematics model indicates multiple motions, i.e., wave-like, straightening, and curling motions can be realized through rotating, twisting, and untwisting of the helical spring, respectively. Through different motions, the robot achieves sidewinding gait, concertina gait, and side-pushing locomotion gait in open space, while lateral and vertical undulations in confined space. Locomotion experiments on floor, on sands, under water, in tubes, and other complex terrain are conducted, which indicates its potential for rescue and exploration missions in complex environments.

Hop Mice Display Synchronous Hindlimb Locomotion and a Ventrally Fused Lumbar Spinal Cord Caused by a Point Mutation in Ttc26.

Identifying the spinal circuits controlling locomotion is critical for unravelling the mechanisms controlling the production of gaits. Development of the circuits governing left-right coordination relies on axon guidance molecules such as ephrins and netrins. To date, no other class of proteins have been shown to play a role during this process. Here, we have analyzed hop mice, which walk with a characteristic hopping gait using their hindlimbs in synchrony. Fictive locomotion experiments suggest that a local defect in the ventral spinal cord contributes to the aberrant locomotor phenotype. Hop mutant spinal cords had severe morphologic defects, including the absence of the ventral midline and a poorly defined border between white and gray matter. The hop mice represent the first model where, exclusively found in the lumbar domain, the left and right components of the central pattern generators (CPGs) are fused with a synchronous hindlimb gait as a functional consequence. These defects were associated with abnormal developmental processes, including a misplaced notochord and reduced induction of ventral progenitor domains. Whereas the underlying mutation in hop mice has been suggested to lie within the Ttc26 gene, other genes in close vicinity have been associated with gait defects. Mouse embryos carrying a CRISPR replicated point mutation within Ttc26 displayed an identical morphologic phenotype. Thus, our data suggest that the assembly of the lumbar CPG network is dependent on fully functional TTC26 protein.

An obstacle-avoiding and stiffness-tunable modular bionic soft robot

AbstractThe aim of this work is to design and model a novel modular bionic soft robot for crawling and crossing obstacles. The modular bionic soft robot is composed of several serial driving soft modules, each module is composed of two parallel soft actuators. By analyzing the influence of working pressure and manufacturing size on the stiffness of the modular bionic soft robot, the nonlinear variable stiffness model of the modular bionic soft robot is established. Based on this model, the spatial states and design parameters of the modular bionic soft robot are discussed when the modular bionic soft robot can pass through the obstacle. Experiments show that when the inflation air pressure of the modular bionic soft robot is 70 kPa, its speed can reach 7.89 mm/s and the height of obstacles passed by it can reach 42.8 mm. The feasibility of the proposed modular bionic soft robot and nonlinear variable stiffness model is verified by locomotion experiments.

Design and Control of a Novel Leg-Arm Multiplexing Mobile Operational Hexapod Robot

A novel legged robot with 6 limbs driven by 20 proprioceptive motors, named SDUHex, is proposed in this letter. The limbs located at the middle of the robot can work as manipulators or locomotors. The topology design of the multi-mode robot is introduced and the kinematics and dynamics of the robot leg is analysed. A hierarchical optimized whole body controller with virtual model active compliance, called VM-WBC, is proposed. This control method unified the foot locomotion and arm operation. The terrain adaptation strategy is proposed to improve the non-flat terrain passability. A series of locomotion and operation experiments are carried out on SDUHex to test the designed robot and control method. Robust locomotion of the robot is verified by lateral impact recovery and slope terrain adaptation experiments. Arm operation of the robot including single-arm pushing and dual-arm picking are experimented to highlight the novel multiplexing function of the robot limbs.

A Novel Dead Reckoning System Based on Wearable Exoskeleton for Rat-Robot Localization

Rat-robots show great potential for rescue and search applications in very tight spaces because of their excellent mobility and environmental adaptability. However, owing to the complexity of the working environment, there is currently no effective localization method. In this study, we present a novel rat-robot localization system based on dead reckoning. A wearable exoskeleton was used to measure the trajectories of the rat-robot’s two hindlimb ankle joints. A relational model was established to estimate the step length from the horizontal distance between the two ankle joints in the sagittal plane, and an inertial sensor obtained the heading of the rat-robot in the navigation coordinate system. The proposed system was tested by both the one-dimensional and two-dimensional locomotion experiment. The step length estimation error of our system was 0.6 ± 3.1%, and the positioning estimation error in the two-dimensional space was 1.65 ± 1.06%. These experimental results demonstrate the efficacy of this system in providing accurate distance and position estimation for the rat-robot.

Design, modeling and control of a miniature bio-inspired amphibious spherical robot

It is a huge challenge for an amphibious robot with a high locomotory performance to locomote in amphibious environments, including crawling on rough terrains, maneuvering underwater, and launching and landing motion between land and water. To deal with such a challenge, a miniature bio-inspired Amphibious Spherical Robot (ASRobot) with a Legged, Multi-vectored Water-jet Composite Driving Mechanism (LMWCDM) has been designed. In this paper, locomotory performance of the robot in amphibious field environments is studied. First, a simplified kinematic model was built to study crawling gaits, and with an online adjustment mechanism, the gaits were adjusted, enabling the robot to climb up slopes more stably. Then, using a dynamic underwater model, a real-time dynamic thrust vectoring allocation strategy is proposed to generate the water-jet thrust and joint angles using desired forces and torques computed by four parallel PID algorithms. Finally, a set of experiments were carried out to evaluate the performance of on land locomotion and underwater locomotion. Further, outdoor locomotion experiments including crawling on various terrains, launching and landing motion, were conducted in field environments. The results demonstrate that the robot prototype possesses the high locomotory performance which endows its wide application of disaster rescue, reconnaissance and resource exploration in amphibious environments.