Biohybrid Control of General Linear Systems Using the Adaptive Filter Model of Cerebellum

Abstract

The adaptive filter model of the cerebellar microcircuit has been successfully applied to biological motor control problems, such as the vestibulo-ocular reflex (VOR), and to sensory processing problems, such as the adaptive cancelation of reafferent noise. It has also been successfully applied to problems in robotics, such as adaptive camera stabilization and sensor noise cancelation. In previous applications to inverse control problems, the algorithm was applied to the velocity control of a plant dominated by viscous and elastic elements. Naive application of the adaptive filter model to the displacement (as opposed to velocity) control of this plant results in unstable learning and control. To be more generally useful in engineering problems, it is essential to remove this restriction to enable the stable control of plants of any order. We address this problem here by developing a biohybrid model reference adaptive control (MRAC) scheme, which stabilizes the control algorithm for strictly proper plants. We evaluate the performance of this novel cerebellar-inspired algorithm with MRAC scheme in the experimental control of a dielectric electroactive polymer, a class of artificial muscle. The results show that the augmented cerebellar algorithm is able to accurately control the displacement response of the artificial muscle. The proposed solution not only greatly extends the practical applicability of the cerebellar-inspired algorithm, but may also shed light on cerebellar involvement in a wider range of biological control tasks.

Visual-tactile sensory map calibration of a biomimetic whiskered robot

“…This method of whisker actuation was incorporated into the platform as part of a broader exploration of cerebellar inspired adaptive control applied to non-linear, time varying plant [9], [10]. The head is attached as the end-effector to an industrial manipulator (ABB -IRB120), hereafter referred to as the neck.…”

“…9. Photograph taken from directly above the head of Bellabot highlighting the intrinsic bend in whisker number 8 that was accommodated into the calibrated map stems from the observation that the reduction in error for the estimates of whisker tip locations of whiskers in the outer circle (9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) were marginally less than the inner circle (1)(2)(3)(4)(5)(6)(7)(8). This has been interpreted as the effective non-linearity introduced by the 2D projection of the 3D whisker tip surface being more pronounced for the more distally located whiskers in the outer circle of the array (see figure 4c).…”

Sparse Bayesian Nonlinear System Identification Using Variational Inference

“…The variational approximation, Q(θ, τ, α) = q(θ, τ )q(α), to the true posterior p(θ, τ, α|y) is found by updating q(θ, τ ) and q(α) by iterating between the update equations defined by (38), (40), (43) and (45), and 47…”

“…The SVB-NARX algorithm was applied in this section to the identification of a set of six dielectric elastomer actuators (DEAs), which are a type of soft-smart actuator used in robotics [36], [48], [49]. DEAs are known to exhibit nonlinear dynamics, posing challenges for control design [37], [38], [52].…”

“…Elements of this work have been published in brief conference format [35], and have been extended here in a number of important ways: a more thorough and rigorous theoretical development, expanded benchmarking, and application to realworld data. The real world data is obtained from a dielectric elastomer actuator (DEA), which is a type of actuator used in soft robotics [36]- [38]. Furthermore, we demonstrate, for the first time, frequency-domain uncertainty analysis for nonlinear systems by propagation of parameter uncertainty from the identified Bayesian model of the DEA into the frequencydomain, using Monte Carlo sampling and nonlinear output frequency response functions (NOFRFs) [6].…”