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

Deep Reinforcement Learning (DRL) has the potential to be used forsynthesizing feedback controllers (agents) for various complex systems withunknown dynamics. These systems are expected to satisfy diverse safety andliveness properties best captured using temporal logic. In RL, the rewardfunction plays a crucial role in specifying the desired behaviour of theseagents. However, the problem of designing the reward function for an RL agentto satisfy complex temporal logic specifications has received limited attentionin the literature. To address this, we provide a systematic way of generatingrewards in real-time by using the quantitative semantics of Signal TemporalLogic (STL), a widely used temporal logic to specify the behaviour ofcyber-physical systems. We propose a new quantitative semantics for STL havingseveral desirable properties, making it suitable for reward generation. Weevaluate our STL-based reinforcement learning mechanism on several complexcontinuous control benchmarks and compare our STL semantics with thoseavailable in the literature in terms of their efficacy in synthesizing thecontroller agent. Experimental results establish our new semantics to be themost suitable for synthesizing feedback controllers for complex continuousdynamical systems through reinforcement learning.

STL-Based Synthesis of Feedback Controllers Using Reinforcement Learning