Real time enhancement of operator's ergonomics in physical human - robot collaboration scenarios using a multi-stereo camera system

2022 IEEE International Conference on Industrial Technology (ICIT), Год журнала: 2023, Номер unknown, С. 1 - 6

Опубликована: Апрель 4, 2023

In collaborative tasks where humans work alongside machines, the robot's movements and behaviour can have a significant impact on operator's safety, health, comfort. To address this issue, we present multi-stereo camera system that continuously monitors posture while they with robot. This uses novel distributed fusion approach to assess in real-time help avoid uncomfortable or unsafe positions. The adjusts informs operator of any incorrect potentially harmful postures, reducing risk accidents, strain, musculoskeletal disorders. analysis is personalized, taking into account unique anthropometric characteristics each operator, ensure optimal ergonomics. results our experiments show proposed leads improved human body postures offers promising solution for enhancing ergonomics operators tasks.

Язык: Английский

---

Future Prospects

Springer eBooks, Год журнала: 2023, Номер unknown, С. 189 - 220

Опубликована: Янв. 1, 2023

Язык: Английский

---

Investigation on the effect of grinding wheel for grinding of AISI D3 tool steel under different conditions

Engineering Research Express, Год журнала: 2022, Номер 4(4), С. 045036 - 045036

Опубликована: Дек. 1, 2022

Abstract The surface finish of ground samples is highly influenced by the grinding parameters, conditions and type wheel. This paper emphasizes on effect various factors such as conditions, wheel operating process parameters like depth cut table speed roughness samples. Two types wheels alumina (Al 2 O 3 ) cubic boron nitride (CBN) were used for AISI D3 tool steel under dry wet conditions. material removal rate evaluated all results showed that outperformed provided a better while using both wheels. Machine Learning was implemented to optimize parameters. Multi-objective optimization genetic algorithm done Pareto frontier chart made help determine what values input would achieve required outputs roughness. different approaches Genetic Algorithm Principle Component Analysis then compared multi-objective optimization. had dominant lesser effect.

Язык: Английский

---

A Novel Application of XAI in Squinting Models: A Position Paper

Опубликована: Янв. 1, 2023

Artificial Intelligence, and Machine Learning especially, are becoming increasingly foundationalto our collective future. Recent developments around generative models such as ChatGPT, andDALL-E represent just the tip of iceberg in new gadgets that will change way we liveour lives. Convolutional Neural Networks (CNNs) Transformer at heart ofadvancements autonomous vehicles health care industries well. Yet these models,as impressive they are, still make plenty mistakes without justifying or explaining whataspects input internal state, was responsible for error. Often, goal automationis to increase throughput, processing many tasks possible a short period time. Forsome use cases cost might be acceptable long production is increased abovesome set margin. However, care, vehicles, financial applications, thecost mistake have catastrophic consequences. For this reason, where singlemistakes can costly less enthusiastic about early AI adoption. The field eXplainable AI(XAI) has attracted significant attention recent years with producing algorithmsthat shed light into decision-making process neural networks. In paper show howrobust vision pipelines built using XAI algorithms automatedwatchdogs actively monitor networks signs ofmistakes ambiguous data. We call robust pipelines, squinting pipelines.

Язык: Английский

---

A novel application of XAI in squinting models: A position paper

Machine Learning with Applications, Год журнала: 2023, Номер 13, С. 100491 - 100491

Опубликована: Авг. 18, 2023

Artificial Intelligence, and Machine Learning especially, are becoming increasingly foundational to our collective future. Recent developments around generative models such as ChatGPT, DALL-E represent just the tip of iceberg in new gadgets that will change way we live lives. Convolutional Neural Networks (CNNs) Transformer at heart advancements autonomous vehicles health care industries well. Yet these models, impressive they are, still make plenty mistakes without justifying or explaining what aspects input internal state, was responsible for error. Often, goal automation is increase throughput, processing many tasks possible a short period time. For some use cases cost might be acceptable long production increased above set margin. However, care, vehicles, financial applications, mistake have catastrophic consequences. this reason, where single can costly less enthusiastic about early AI adoption. The field eXplainable (XAI) has attracted significant attention recent years with producing algorithms shed light into decision-making process neural networks. In paper show how robust vision pipelines built using XAI automated watchdogs actively monitor networks signs ambiguous data. We call pipelines, squinting pipelines.

Язык: Английский

---