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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Reinforcement Learning (RL) agents trained in simulated environments and thendeployed in the real world are often sensitive to the differences in dynamicspresented, commonly termed the sim-to-real gap. With the goal of minimizingthis gap on resource-constrained embedded systems, we train and live-adaptagents on quadrotors built from off-the-shelf hardware. In achieving this wedeveloped three novel contributions. (i) SwaNNFlight, an open-source firmwareenabling wireless data capture and transfer of agents' observations.Fine-tuning agents with new data, and receiving and swapping onboard NNcontrollers -- all while in flight. We also design SwaNNFlight System (SwaNNFS)allowing new research in training and live-adapting learning agents on similarsystems. (ii) Multiplicative value composition, a technique for preserving theimportance of each policy optimization criterion, improving trainingperformance and variability in learnt behavior. And (iii) anchor critics tohelp stabilize the fine-tuning of agents during sim-to-real transfer, onlinelearning from real data while retaining behavior optimized in simulation. Wetrain consistently flight-worthy control policies in simulation and deploy themon real quadrotors. We then achieve live controller adaptation via over-the-airupdates of the onboard control policy from a ground station. Our resultsindicate that live adaptation unlocks a near-50\% reduction in powerconsumption, attributed to the sim-to-real gap. Finally, we tackle the issuesof catastrophic forgetting and controller instability, showing theeffectiveness of our novel methods.  Project Website: https://github.com/BU-Cyber-Physical-Systems-Lab/SwaNNFS

The SwaNNFlight System: On-the-Fly Sim-to-Real Adaptation via Anchored Learning