Squeezing More Past Knowledge for Online Class-Incremental Continual Learning

Continual learning(CL)studies the problem of learning to accumulate knowledge over time from a stream of data.A crucial challenge is that neural networks suffer from performance degradation on previously seen data,known as catastrophic forgetting,due to allowing parameter sharing.In this work,we consider a more practical online class-incremental CL setting,where the model learns new samples in an online manner and may continuously experience new classes.Moreover,prior knowledge is unavailable during training and evaluation.Existing works usually explore sample usages from a single dimension,which ignores a lot of valuable supervisory information.To better tackle the setting,we propose a novel replay-based CL method,which leverages multi-level representations produced by the intermediate process of training samples for replay and strengthens supervision to consolidate previous knowledge.Specifically,besides the previous raw samples,we store the corresponding logits and features in the memory.Furthermore,to imitate the prediction of the past model,we construct extra constraints by leveraging multi-level information stored in the memory.With the same number of samples for replay,our method can use more past knowledge to prevent interference.We conduct extensive evaluations on several popular CL datasets,and experiments show that our method consistently outperforms state-of-the-art methods with various sizes of episodic memory.We further provide a detailed analysis of these results and demonstrate that our method is more viable in practical scenarios.

Reinforcement Learning Navigation for Robots Based on Hippocampus Episode Cognition

Artificial intelligence is currently achieving impressive success in all fields.However,autonomous navigation remains a major challenge for AI.Reinforcement learning is used for target navigation to simulate the interaction between the brain and the environment at the behavioral level,but the Artificial Neural Network trained by reinforcement learning cannot match the autonomous mobility of humans and animals.The hippocampus–striatum circuits are considered as key circuits for target navigation planning and decision-making.This paper aims to construct a bionic navigation model of reinforcement learning corresponding to the nervous system to improve the autonomous navigation performance of the robot.The ventral striatum is considered to be the behavioral evaluation region,and the hippocampal–striatum circuit constitutes the position–reward association.In this paper,a set of episode cognition and reinforcement learning system simulating the mechanism of hippocampus and ventral striatum is constructed,which is used to provide target guidance for the robot to perform autonomous tasks.Compared with traditional methods,this system reflects the high efficiency of learning and better Environmental Adaptability.Our research is an exploration of the intersection and fusion of artificial intelligence and neuroscience,which is conducive to the development of artificial intelligence and the understanding of the nervous system.

Bio-Inspired Controller for a Robot Cheetah with a Neural Mechanism Controlling Leg Muscles

The realization of a high-speed running robot is one of the most challenging problems in developing legged robots. The excellent performance of cheetahs provides inspiration for the control and mechanical design of such robots. This paper presents a three-dimensional model of a cheetah that predicts the locomotory behaviors of a running cheetah. Applying biological knowledge of the neural mechanism, we control the muscle flexion and extension during the stance phase, and control the positions of the joints in the flight phase via a PD controller to minimize complexity. The proposed control strategy is shown to achieve similar locomotion of a real cheetah. The simulation realizes good biological properties, such as the leg retraction, ground reaction force, and spring-like leg behavior. The stable bounding results show the promise of the controller in high-speed locomotion. The model can reach 2.7 m-s^-1 as the highest speed, and can accelerate from 0 to 1.5 m-s^-1 in one stride cycle. A mechanical structure based on this simulation is designed to demonstrate the control approach, and the most recently developed hindlimb controlled by the proposed controller is presented in swinging-leg experiments and jump-force experiments.

Approximate Perturbation Stance Map of the SLIP Runner and Application to Locomotion Control

This paper presents a novel method of perturbation to obtain the analytic approximate solution to the Spring-Loaded Inverted Pendulum （SLIP） dynamics in stance phase with considering the effect of gravity. This perturbation solution achieves higher accuracy in predicting the apex state variables than the typical existing analytic approximations. Particularly, our solution is validated for non-symmetric trajectory of hopping in a large angle range. Furthermore, the stance controller of the SLIP runner is developed to regulate the apex state based on the approximate apex return map. To compensate the energy variation between the current and desired apex states, a stiffness adjustment of the leg spring in stance phase is presented. The deadbeat controller of the angle of attack is designed to track the regulated apex height and velocity. The simulation demonstrates that the SLIP runner applying the proposed stance controller reveals higher tracking accuracy and more rapidly converges to the regulated apex state.

Force-controlled Compensation Scheme for P-Q Valve-controlled Asymmetric Cylinder used on Hydraulic Quadruped Robots

Under the requirement of the force controller of hydraulic quadruped robots,the goal of this work is to accurately track the force commands at the level of the hydraulic drive unit.The main contribution focuses on the development of a force-controlled compensation scheme,which is specifically aimed at the key issues affecting the hydraulic quadrupedal locomotion.With this idea,based on a P-Q valve-controlled asymmetric cylinder,we first establish a mathematical model for the hydraulic drive unit force control system.With the desired force commands,a force feed-forward algorithm is presented to improve the dynamic performance of the system.Meanwhile,we propose a disturbance compensation algorithm to reduce the influence induced by external disturbances due to foot-ground impacts.Afterwards,combining with a variable gain PI controller,a series of experiments are implemented on a force control performance test platform to verify the proposed scheme.The results demonstrate that the force-controlled compensation scheme has the ability to notably improve the force tracking accuracy,reduce the response time and redundant force.

A Unified Trajectory Optimization Approach for Long-Term and Reactive Motion Planning of Legged Locomotion

This paper proposes a unified trajectory optimization approach that simultaneously optimizes the trajectory of the center of mass and footholds for legged locomotion.Based on a generic point-mass model,the approach is formulated as a nonlinear optimization problem,incorporating constraints such as robot kinematics,dynamics,ground reaction forces,obstacles,and target location.The unified optimization approach can be applied to both long-term motion planning and the reactive online planning through the use of model predictive control,and it incorporates vector field guidance to converge to the long-term planned motion.The effectiveness of the approach is demonstrated through simulations and physical experiments,showing its ability to generate a variety of walking and jumping gaits,as well as transitions between them,and to perform reactive walking in obstructed environments.

Robot Navigation Strategy in Complex Environment Based on Episode Cognition

The hippocampal formation of the brain contains a series of nerve cells related to environmental cognition and navigation.These cells can integrate their moment information and external perceptual information and acquire episodic cognitive memory.Through episodic cognition and memory,organisms can achieve autonomous navigation in complex environments.This paper mainly studies the strategy of robot episode navigation in complex environments.After exploring the environment,the robot obtains subjective environmental cognition and forms a cognition map.The grid cells information contained in the cognitive map can obtain the direction and distance of the target through vector calculation,which can get a shortcut through the inexperienced area.The synaptic connection of place cells in the cognitive map can be used as the topological relationship between episode nodes.When the target-oriented vector navigation encounters obstacles,the obstacles can be realized by setting closer sub-targets.Based on the known obstacle information obtained from boundary cells in the cognitive map,topological paths can be divided into multi-segment vector navigation to avoid encountering obstacles.This paper combines vector and topological navigation to achieve goal-oriented and robust navigation capability in a complex environment.

Control of a Quadruped Robot with Bionic Springy Legs in Trotting Gait

Legged robots have better performance on discontinuous terrain than that of wheeled robots. However, the dynamic trotting and balance control of a quadruped robot is still a challenging problem, especially when the robot has multi-joint legs. This paper presents a three-dimensional model of a quadruped robot which has 6 Degrees of Freedom （DOF） on torso and 5 DOF on each leg. On the basis of the Spring-Loaded Inverted Pendulum （SLIP） model, body control algorithm is discussed in the first place to figure out how legs work in 3D trotting. Then, motivated by the principle of joint function separation and introducing certain biological characteristics, two joint coordination approaches are developed to produce the trot and provide balance. The robot reaches the highest speed of 2.0 m.s-1, and keeps balance under 250 Kg.m.s-1 lateral disturbance in the simulations. The effectiveness of these approaches is also verified on a prototype robot which runs to 0.83 m.s-1 on the treadmill, The simulations and experiments show that legged robots have good biological properties, such as the ground reaction force, and spring-like leg behavior.

CPG Control for Biped Hopping Robot in Unpredictable Environment

A CPG control mechanism is proposed for hopping motion control of biped robot in unpredictable environment. Based on analysis of robot motion and biological observation of animal＇s control mechanism, the motion control task is divided into two simple parts： motion sequence control and output force control. Inspired by a two-level CPG model, a two-level CPG control mechanism is constructed to coordinate the drivers of robot joint, while various feedback information are introduced into the control mechanism. Interneurons within the control mechanism are modeled to generate motion rhythm and pattern promptly for motion sequence control; motoneurons are modeled to control output forces of joint drivers in real time according to feedbacks. The control system can perceive changes caused by unknown perturbations and environment changes according to feedback information, and adapt to unpredictable environment by adjusting outputs of neurons. The control mechanism is applied to a biped hopping robot in unpredictable environment on simulation platform, and stable adaptive motions are obtained.

Velocity Control of a Bounding Quadruped via Energy Control and Vestibular Reflexes

In this paper a bio-inspired approach of velocity control for a quadruped robot running with a bounding gait on compliant legs is set up. The dynamic properties ofa sagittal plane model of the robot are investigated. By analyzing the stable fixed points based on Poincare map, we find that the energy change of the system is the main source for forward velocity adjustment. Based on the analysis of the dynamics model of the robot, a new simple linear running controller is proposed using the energy control idea, which requires minimal task level feedback and only controls both the leg torque and ending impact angle. On the other hand, the functions of mammalian vestibular reflexes are discussed, and a reflex map between forward velocity and the pitch movement is built through statistical regression analysis. Finally, a velocity controller based on energy control and vestibular reflexes is built, which has the same structure as the mammalian nervous mechanism for body posture control. The new con- troller allows the robot to run autonomously without any other auxiliary equipment and exhibits good speed adjustment capa- bility. A series simulations and experiments were set to show the good movement agility, and the feasibility and validity of the robot system.

A Parallel Actuated Pantograph Leg for High-speed Locomotion

High-speed running is one of the most important topics in the field of legged robots which requires strict constraints on structural design and control. To solve the problems of high acceleration, high energy consumption, high pace frequency and ground impact during high-speed movement, this paper presents a parallel actuated pantograph leg with an approximately decoupled configuration. The articulated leg features in light weight, high load capacity, high mechanical efficiency and structural stability. The similarity features of force and position between the control point and the foot are analyzed. The key design parameters, K1 and K2, which concern the dynamic performances, are carefully optimized by comprehensive evaluation of the leg inertia and mass within the maximum foot trajectory, A control strategy that incorporates virtual Spring Loaded Inverted Pendulum （SLIP） model and active force is also proposed to test the design. The strategy can implement highly flexible impedance without mechanical springs, which substantially simplifies the design and satisfies the variable stiffness requirements during high-speed running. The rationality of the structure and the effectiveness of the control law are validated by simulation and experiments.