scholarly journals Robust assembly sequence generation in a Human-Robot Collaborative workcell by reinforcement learning

2021 ◽  
Vol 49 (4) ◽  
pp. 851-858
Author(s):  
Dario Antonelli ◽  
Qingfei Zeng ◽  
Khurshid Aliev ◽  
Xuemei Liu
Keyword(s):  
Reinforcement Learning ◽  
Cognitive Effort ◽  
Robot Programming ◽  
Task Sequence ◽  
Collaborative Robots ◽  
Sequence Generation ◽  
Industrial Assembly ◽  
Physical Limitations ◽  
Assembly Sequence Generation

Human-Robot Collaborative (HRC) workcells could enhance the inclusive employment of human workers regardless their force or skills. Collaborative robots not only substitute humans in dangerous and heavy tasks, but also make the related processes within the reach of all workers, overcoming lack of skills and physical limitations. To enable the full exploitation of collaborative robots traditional robot programming must be overcome. Reduction of robot programming time and worker cognitive effort during the job become compelling requirements to be satisfied. Reinforcement learning (RL) plays a core role to allow robot to adapt to a changing and unstructured environment and to human undependable execution of repetitive tasks. The paper focuses on the utilization of RL to allow a robust industrial assembly process in a HRC workcell. The result of the study is a method for the online generation of robot assembly task sequence that adapts to the unpredictable and inconstant behavior of the human co-workers. The method is presented with the help of a benchmark case study.

Deep Reinforcement Learning for Multiparameter Optimization in de novo Drug Design

2019 ◽  
Author(s):  
Niclas Ståhl ◽  
Göran Falkman ◽  
Alexander Karlsson ◽  
Gunnar Mathiason ◽  
Jonas Boström
Keyword(s):  
Reinforcement Learning ◽  
Short Term Memory ◽  
De Novo ◽  
De Novo Drug Design ◽  
Generative Process ◽  
New Methods ◽  
Multiparameter Optimization ◽  
Long Short Term Memory ◽  
New Compounds

<p>In medicinal chemistry programs it is key to design and make compounds that are efficacious and safe. This is a long, complex and difficult multi-parameter optimization process, often including several properties with orthogonal trends. New methods for the automated design of compounds against profiles of multiple properties are thus of great value. Here we present a fragment-based reinforcement learning approach based on an actor-critic model, for the generation of novel molecules with optimal properties. The actor and the critic are both modelled with bidirectional long short-term memory (LSTM) networks. The AI method learns how to generate new compounds with desired properties by starting from an initial set of lead molecules and then improve these by replacing some of their fragments. A balanced binary tree based on the similarity of fragments is used in the generative process to bias the output towards structurally similar molecules. The method is demonstrated by a case study showing that 93% of the generated molecules are chemically valid, and a third satisfy the targeted objectives, while there were none in the initial set.</p>

End-User Robot Programming Case Study

2020 ◽  
Author(s):  
Michal Kapinus ◽  
Zdeněk Materna ◽  
Daniel Bambušek ◽  
Vitězslav Beran
Keyword(s):  
Robot Programming ◽  
End User

Solution space reduction of disassembly sequences generated automatically via computer aids

2014 ◽  
Vol 229 (16) ◽  
pp. 2977-2986 ◽  
Author(s):  
Louisa Issaoui ◽  
Nizar Aifaoui ◽  
Abdelmajid Benamara
Keyword(s):  
Solution Space ◽  
Target Component ◽  
Cad Model ◽  
Simulation Tool ◽  
Sequence Generation ◽  
Space Solution ◽  
Space Reduction ◽  
Automatic Sequence ◽  
Selective Disassembly

To develop a simulation tool for automatic disassembly in computer aid design (CAD) environment two difficulties are found: the huge space of generated sequences and their feasibility especially in combinatory generation. This article deals with automatic sequence generation for selective disassembly of mechanical product. Starting from a CAD model a new appropriate connection tree of a target component is constructed. This tree aims at reducing space solution by eliminatory rules. The generation of sequence is based on reading of connection tree branches and eliminatory tests of feasibility. The feasibility is checked by updating the disassembly mobility of each sequence’s element. A case study is presented to prove the effectiveness of the proposed approach.

A Reinforcement Learning - Great-Deluge Hyper-Heuristic for Examination Timetabling

2010 ◽  
Vol 1 (1) ◽  
pp. 39-59 ◽  
Author(s):  
Ender Özcan ◽  
Mustafa Misir ◽  
Gabriela Ochoa ◽  
Edmund K. Burke
Keyword(s):  
Reinforcement Learning ◽  
Complete Solution ◽  
Examination Timetabling ◽  
Low Level ◽  
Termination Criteria ◽  
Candidate Solution ◽  
Wide Range ◽  
Finite Set ◽  
Different Characteristics

Hyper-heuristics can be identified as methodologies that search the space generated by a finite set of low level heuristics for solving search problems. An iterative hyper-heuristic framework can be thought of as requiring a single candidate solution and multiple perturbation low level heuristics. An initially generated complete solution goes through two successive processes (heuristic selection and move acceptance) until a set of termination criteria is satisfied. A motivating goal of hyper-heuristic research is to create automated techniques that are applicable to a wide range of problems with different characteristics. Some previous studies show that different combinations of heuristic selection and move acceptance as hyper-heuristic components might yield different performances. This study investigates whether learning heuristic selection can improve the performance of a great deluge based hyper-heuristic using an examination timetabling problem as a case study.

Sign in / Sign up

Export Citation Format

Share Document