Efficient stochastic model predictive control for embedded systems based on second-order cone programs
| Autor: | Sergio Lucia, Pablo Zometa, Hannes Heinemann, Rolf Findeisen, Markus Kögel |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
0209 industrial biotechnology
Mathematical optimization 021103 operations research Polynomial chaos business.industry Reliability (computer networking) Embedded hardware 0211 other engineering and technologies Stochastic model predictive control 02 engineering and technology Cone (formal languages) Microcontroller Model predictive control 020901 industrial engineering & automation Control theory Order (exchange) Embedded system business Mathematics |
| Zdroj: | ECC |
| Popis: | While by now efficient formulations and tailored solution approaches for nominal model predictive control exist, results for the uncertain case on embedded systems are less common. We consider stochastic model predictive control based on a polynomial chaos expansion formulation. We describe a tailored structure-exploiting solution approach for the resulting second-order cone program. We outline techniques that improve the approach's numerical reliability on embedded hardware with low numerical precision. The efficiency and performance of the approach is experimentally validated on a microcontroller. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|