File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1299/JAMDSM.2021JAMDSM0007
- Scopus: eid_2-s2.0-85101046527
- WOS: WOS:000744080800004
Supplementary
- Citations:
- Appears in Collections:
Article: Bayesian heterogeneous assembly time modeling for robotic performance prediction
Title | Bayesian heterogeneous assembly time modeling for robotic performance prediction |
---|---|
Authors | |
Keywords | Bayesian modeling Gibbs sampler Hazard regression Population heterogeneity Process improvement |
Issue Date | 2021 |
Citation | Journal of Advanced Mechanical Design, Systems and Manufacturing, 2021, v. 15, n. 1 How to Cite? |
Abstract | Robotic systems are widely applied in manufacturing industry to reduce labor cost and increase productivity. Due to the influence of different observed factors (e.g., controllable robotic settings) and the unobserved heterogeneity caused by the unobserved/unknown factors (e.g., material non-uniformity, geometry variability of assembly units) during the assembly process, the assembly time of assembly units exhibits highly heterogeneous performance. To improve the productivity of the assembly process, it is important to develop an analytics-based model to provide accurate assembly time prediction of assembly units and further identify the potential influencing factors for further process improvement. This paper proposes a full Bayesian modeling approach to predicting the performance of heterogeneous robotic assembly time data. Specifically, a generic statistical model is provided to characterize the heterogeneous assembly time data of a heterogeneous population of assembly units by considering the influence of both observed factors as well as unobserved heterogeneity. Bayesian sampling algorithm is further developed to address a series of parameters estimation challenges, such as highly correlation among parameters from different sub-populations and non-conjugate priors. With the developed Bayesian estimation algorithm, both the influence of observed factors as well unobserved heterogeneity can be jointly quantified with both point estimates and exact internal estimates obtained. Both the numerical and case studies are further provided to justify the validity of the proposed modeling approach and demonstrate its superior prediction performance. The proposed prediction model with considering the population heterogeneity of assembly units exhibits better assembly time prediction performance as compared to alternative modeling approaches with the restrictive assumption of population homogeneity. Moreover, the proposed model is able to identify the relevant factors and quantify their influence with parameters uncertainty quantification for further performance improvement of the assembly process. |
Persistent Identifier | http://hdl.handle.net/10722/330433 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Li, Mingyang | - |
dc.contributor.author | Sun, Xuxue | - |
dc.contributor.author | Liang, Guoyuan | - |
dc.contributor.author | Shen, Yingjun | - |
dc.contributor.author | Zhang, Qingpeng | - |
dc.contributor.author | Feng, Yachun | - |
dc.date.accessioned | 2023-09-05T12:10:35Z | - |
dc.date.available | 2023-09-05T12:10:35Z | - |
dc.date.issued | 2021 | - |
dc.identifier.citation | Journal of Advanced Mechanical Design, Systems and Manufacturing, 2021, v. 15, n. 1 | - |
dc.identifier.uri | http://hdl.handle.net/10722/330433 | - |
dc.description.abstract | Robotic systems are widely applied in manufacturing industry to reduce labor cost and increase productivity. Due to the influence of different observed factors (e.g., controllable robotic settings) and the unobserved heterogeneity caused by the unobserved/unknown factors (e.g., material non-uniformity, geometry variability of assembly units) during the assembly process, the assembly time of assembly units exhibits highly heterogeneous performance. To improve the productivity of the assembly process, it is important to develop an analytics-based model to provide accurate assembly time prediction of assembly units and further identify the potential influencing factors for further process improvement. This paper proposes a full Bayesian modeling approach to predicting the performance of heterogeneous robotic assembly time data. Specifically, a generic statistical model is provided to characterize the heterogeneous assembly time data of a heterogeneous population of assembly units by considering the influence of both observed factors as well as unobserved heterogeneity. Bayesian sampling algorithm is further developed to address a series of parameters estimation challenges, such as highly correlation among parameters from different sub-populations and non-conjugate priors. With the developed Bayesian estimation algorithm, both the influence of observed factors as well unobserved heterogeneity can be jointly quantified with both point estimates and exact internal estimates obtained. Both the numerical and case studies are further provided to justify the validity of the proposed modeling approach and demonstrate its superior prediction performance. The proposed prediction model with considering the population heterogeneity of assembly units exhibits better assembly time prediction performance as compared to alternative modeling approaches with the restrictive assumption of population homogeneity. Moreover, the proposed model is able to identify the relevant factors and quantify their influence with parameters uncertainty quantification for further performance improvement of the assembly process. | - |
dc.language | eng | - |
dc.relation.ispartof | Journal of Advanced Mechanical Design, Systems and Manufacturing | - |
dc.subject | Bayesian modeling | - |
dc.subject | Gibbs sampler | - |
dc.subject | Hazard regression | - |
dc.subject | Population heterogeneity | - |
dc.subject | Process improvement | - |
dc.title | Bayesian heterogeneous assembly time modeling for robotic performance prediction | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1299/JAMDSM.2021JAMDSM0007 | - |
dc.identifier.scopus | eid_2-s2.0-85101046527 | - |
dc.identifier.volume | 15 | - |
dc.identifier.issue | 1 | - |
dc.identifier.eissn | 1881-3054 | - |
dc.identifier.isi | WOS:000744080800004 | - |