Robotic Action-state Evaluation via Siamese Neural Network

Research output: Contribution to conferencePaperpeer-review

17 Downloads (Pure)


Robotic imitation learning methods assist robots to operate in evolving and unconstrained environments. However, current robotic state representation imitation learning methods still must involve human experts to provide sparse rewards that indicate whether robots successfully complete tasks. However, enabling robots to make the action-state evaluation autonomously still remains a challenge, especially for multi-stage complex tasks. Therefore, in this work, we propose a novel Siamese neural network-based robotic action state evaluation system in an imitation learning system, so as to replace human experts in a multi-stage imitation learning process and improve the learning efficiency. One target learning footage is divided into several stages; for each stage, two Siamese network frameworks are created to assess the robotic action-states in terms of both movement and environment changes.
Original languageEnglish
Publication statusPublished - 2022
EventRobotics for Unconstrained Environments: Robotics for Unconstrained Environments - Aberystwyth, Aberystwyth, United Kingdom of Great Britain and Northern Ireland
Duration: 25 Aug 202226 Aug 2022
Conference number: 5


ConferenceRobotics for Unconstrained Environments
Abbreviated titleUK-RAS 2022
Country/TerritoryUnited Kingdom of Great Britain and Northern Ireland
Period25 Aug 202226 Aug 2022
OtherThe UKRAS conference series has an impor-
tant aim of inclusion to allow young researchers
to present their work and share ideas with peers
and more senior researchers. As such, this year’s
conference attempted to offer as many opportunities
for discussions.
The conference also provided a platform to high-
light UK robotics research in the area of uncon-
strained environments.
Internet address


  • robotic action state evaluation
  • siamese neural network
  • imitation learning
  • few-shot learning


Dive into the research topics of 'Robotic Action-state Evaluation via Siamese Neural Network'. Together they form a unique fingerprint.

Cite this