A Cerebellum-Inspired Learning Approach for Adaptive and Anticipatory Control

Abstract

The cerebellum, which is responsible for motor control and learning, has been suggested to act as a Smith predictor for compensation of time-delays by means of internal forward models. However, insights about how forward model predictions are integrated in the Smith predictor have not yet been unveiled. To fill this gap, a novel bio-inspired modular control architecture that merges a recurrent cerebellar-like loop for adaptive control and a Smith predictor controller is proposed. The goal is to provide accurate anticipatory corrections to the generation of the motor commands in spite of sensory delays and to validate the robustness of the proposed control method to input and physical dynamic changes. The outcome of the proposed architecture with other two control schemes that do not include the Smith control strategy or the cerebellar-like corrections are compared. The results obtained on four sets of experiments confirm that the cerebellum-like circuit provides more effective corrections when only the Smith strategy is adopted and that minor tuning in the parameters, fast adaptation and reproducible configuration are enabled.

Online Learning of the Dynamical Internal Model of Transfemoral Prosthesis for Enhancing Compliance

“…The proposed control architecture is bio-inspired, i.e., (i) it implements artificial Central Pattern Generators (CPGs) using an adaptive oscillator [37], [38], and (ii) it partially mimics the function of the cerebellum in human locomotion. Indeed, previous studies on human motor control highlighted the key role of the cerebellum in motor control and learning, and support the assumption that this neural structure encodes forward and inverse internal models [39] to achieve accurate and coordinated movements [40]. In this way, the putative role of the cerebellum is twofold: (1) it provides a feedforward predictive action based on continuous learning of the task, and (2) it provides an error-correction action to ensure robustness to perturbations [41].…”

A multizone cerebellar chip for bioinspired adaptive robot control and sensorimotor processing

“…Previous work in robotics has used cerebellar-inspired algorithms to provide adaptive solutions for robot control in single applications including variable stiffness, lightweight actuators with varying dynamics [13][14][15] especially in the context of robot arm control [16][17][18][19][20][21][22]. Cerebellar inspired models have also been applied in robotics to locomotion [23], collision or obstacle avoidance tasks [19,[24][25][26][27], to gaze stabilization tasks [19,28,29], to the adaptive cancellation of self-generated sensory signals [30], and to provide anticipatory control [31]. However, cerebellarinspired control algorithms have not yet been tested through simultaneous application of the same microzone algorithm to a range of different tasks within a single robotic system.…”

A Fully Spiking Neural Control System Based on Cerebellar Predictive Learning for Sensor-Guided Robots

“…Cerebellar models were categorized in [11], which they can be either a state-encoder-driven model, functional or cellular model. More attention is given to functional models to approximate functions executed by each cerebellar layer, which makes it more appealing from computer-science perspective [9]. While the cellular models are more challenging, it is not crucial only to verify and develop theories about the learning mechanisms, but to gain insight as well on how neurons act on individual and population basis to achieve such task [8].…”

“…The SNNs incorporate most of the biological neuronal dynamics which include complex and realistic firing patterns based on spikes [14]. Besides, the novelty of the work is that the forward cerebellar predictions are integrated as a Smith predictor [9] and they are supervised by the error in task space.…”