Dynamic Offloading of Control Algorithms to the Edge using 5G and WebAssembly
(2024)Department of Automatic Control
- Abstract
- The aim of this work was to test if WebAssembly universal byte code and 5G communication technology is suitable in the context of offloading control mission of real-time systems. To test these tools a new dynamic offloading framework was implemented and tested on a Furuta pendulum, an inherently unstable and time-critical process using, among other computing units, an edge node as the offloading target. The implementation is considered dynamic because: 1) The local device which interacts with the I/O of the process dynamically send the control application to be used by the edge node. 2) The local device dynamically decides on which controller should control the process, either the local fallback linear quadratic regulator controller or the... (More)
- The aim of this work was to test if WebAssembly universal byte code and 5G communication technology is suitable in the context of offloading control mission of real-time systems. To test these tools a new dynamic offloading framework was implemented and tested on a Furuta pendulum, an inherently unstable and time-critical process using, among other computing units, an edge node as the offloading target. The implementation is considered dynamic because: 1) The local device which interacts with the I/O of the process dynamically send the control application to be used by the edge node. 2) The local device dynamically decides on which controller should control the process, either the local fallback linear quadratic regulator controller or the CVXGEN Model Predictive Control solver written in an Ahead-of-Time WebAssembly format, which is used by the edge node. The work concluded that Wasm run in interpreted form was too slow to control the process but AOT compiled Wasm which could be run by WasmEdge runtime worked well with an execution time close to the native speed. It was also concluded that data transfer using 5G technology without uRLCC was fast enough to balance the pendulum and is suitable for offloading, other communication medium where also tested in this work such as: Wi-Fi and cabled Ethernet. In the study we also developed a control quality measure for decision making on when to offload. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9153618
- author
- Al Bayati, Ahmed
- supervisor
-
- Karl-Erik Årzén LU
- Johan Eker LU
- organization
- year
- 2024
- type
- H3 - Professional qualifications (4 Years - )
- subject
- report number
- TFRT-6231
- other publication id
- 0280-5316
- language
- English
- id
- 9153618
- date added to LUP
- 2024-05-22 16:10:45
- date last changed
- 2024-05-22 16:10:45
@misc{9153618, abstract = {{The aim of this work was to test if WebAssembly universal byte code and 5G communication technology is suitable in the context of offloading control mission of real-time systems. To test these tools a new dynamic offloading framework was implemented and tested on a Furuta pendulum, an inherently unstable and time-critical process using, among other computing units, an edge node as the offloading target. The implementation is considered dynamic because: 1) The local device which interacts with the I/O of the process dynamically send the control application to be used by the edge node. 2) The local device dynamically decides on which controller should control the process, either the local fallback linear quadratic regulator controller or the CVXGEN Model Predictive Control solver written in an Ahead-of-Time WebAssembly format, which is used by the edge node. The work concluded that Wasm run in interpreted form was too slow to control the process but AOT compiled Wasm which could be run by WasmEdge runtime worked well with an execution time close to the native speed. It was also concluded that data transfer using 5G technology without uRLCC was fast enough to balance the pendulum and is suitable for offloading, other communication medium where also tested in this work such as: Wi-Fi and cabled Ethernet. In the study we also developed a control quality measure for decision making on when to offload.}}, author = {{Al Bayati, Ahmed}}, language = {{eng}}, note = {{Student Paper}}, title = {{Dynamic Offloading of Control Algorithms to the Edge using 5G and WebAssembly}}, year = {{2024}}, }