English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The performance of robots in high-level tasks depends on the quality of theirlower-level controller, which requires fine-tuning. However, the intrinsicallynonlinear dynamics and controllers make tuning a challenging task when it isdone by hand. In this paper, we present DiffTune, a novel, gradient-basedautomatic tuning framework. We formulate the controller tuning as a parameteroptimization problem. Our method unrolls the dynamical system and controller asa computational graph and updates the controller parameters throughgradient-based optimization. The gradient is obtained using sensitivitypropagation, which is the only method for gradient computation when tuning fora physical system instead of its simulated counterpart. Furthermore, we use$\mathcal{L}_1$ adaptive control to compensate for the uncertainties (thatunavoidably exist in a physical system) such that the gradient is not biased bythe unmodelled uncertainties. We validate the DiffTune on a Dubin's car and aquadrotor in challenging simulation environments. In comparison withstate-of-the-art auto-tuning methods, DiffTune achieves the best performance ina more efficient manner owing to its effective usage of the first-orderinformation of the system. Experiments on tuning a nonlinear controller forquadrotor show promising results, where DiffTune achieves 3.5x tracking errorreduction on an aggressive trajectory in only 10 trials over a 12-dimensionalcontroller parameter space.

DiffTune: Auto-Tuning through Auto-Differentiation