ﺑﻬﯿﻨﻪ ﺳﺎزي اﻓﻖﻫﺎي ﮐﻨﺘﺮل ﭘﯿﺶ ﺑﯿﻦ ﻣﺪل ﺑﺎ اﺳﺘﻔﺎده از اﻟﮕﻮرﯾﺘﻢ ازدﺣﺎم ذرات در راﺳﺘﺎي ﻫﻢﮔﺎمﺳﺎزي ﺣﺮﮐﺖ ﺷﺒﯿﻪ ﺳﺎز درﯾﺎﯾﯽ

Optimization of Model Predictive Control Horizons Using Particle Swarm Algorithm to Synchronize Marine Simulator Motion

ﺷﺒﯿﻪﺳﺎزﻫﺎي درﯾﺎﯾﯽ، اﺑﺰارﻫﺎي ﻣﺆﺛﺮي ﺑﺮاي اﺣﺴﺎس راﻧﺪن ﯾﮏ ﺷﻨﺎور درﯾﺎﯾﯽ از ﻃﺮﯾﻖ اﯾﺠﺎد ﯾﮏ ﻣﺤﯿﻂ ﻣﺸﺎﺑﻪ ﺑﺎ اﺳﺘﻔﺎده از ﻓﺮﻣﺎنﻫﺎي ﺣﺮﮐﺘﯽ ﻫﺴﺘﻨﺪ. ﻣﺸﮑﻞ اﺻﻠﯽ ﺷﺒﯿﻪﺳﺎزﻫﺎ ﻓﻀﺎي ﮐﺎر ﻣﺤﺪود ي اﺳﺖ ﮐﻪ ﺑﻪ آنﻫﺎ اﺟﺎزه ﻧﻤﯽ دﻫﺪ ﺗﺎ ﺣﺮﮐﺎت دﻗﯿﻖ ﺷﻨﺎور واﻗﻌﯽ را اﯾﺠﺎد ﮐﻨﻨﺪ؛ در ﻧﺘﯿﺠﻪ آنﻫﺎ ﺑﻪ اﻟﮕﻮرﯾﺘﻢ ﻫﻢ ﮔﺎمﺳﺎزي ﺣﺮﮐﺖ ﻧﯿﺎز دارﻧﺪ. اﺧﯿﺮاً اﺳﺘﻔﺎده از ﮐﻨﺘﺮل ﭘﯿﺶﺑﯿﻦ در ﺷﺒﯿﻪﺳﺎزﻫﺎي درﯾﺎﯾﯽ ﺑﻪ ﻣﺤﺒﻮﺑﯿﺖ رﺳﯿﺪه اﺳﺖ. درﯾﭽﻪﻫﺎي اﻓﻖ ﮐﻨﺘﺮل و ﭘﯿﺶﺑﯿﻨﯽ آﯾﻨﺪه ﺑﺮ ﺑﺎر ﻣﺤﺎﺳﺒﺎﺗﯽ ﺗﺄﺛﯿﺮ ﻣﯽﮔﺬارد اﻣﺎ از آن ﺟﺎ ﮐﻪ اﯾﻦ اﻓﻖﻫﺎ ﺑﻪﺻﻮرت دﺳﺘﯽ ﺗﻮﺳﻂ ﻃﺮاح اﻧﺘﺨﺎب ﻣﯽﺷﻮﻧﺪ، ﭘﺎﯾﯿﻦﺗﺮ از ﺳﻄﺢ ﺑﻬﯿﻨﻪ ﻣﯽﺑﺎﺷﻨﺪ. در اﯾﻦ ﻣﻘﺎﻟﻪ، روﺷﯽ ﻧﻮﯾﻦ ﺑﺮ ﻣﺒﻨﺎي اﻟﮕﻮرﯾﺘﻢ ازدﺣﺎم ذرات ﺑﺮاي دﺳﺘﯿﺎﺑﯽ ﺑﻪ ﺑﻬﺘﺮﯾﻦ اﻓﻖﻫﺎي ﮐﻨﺘﺮل و ﭘﯿﺶﺑﯿﻨﯽ ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﺣﺪاﻗﻞرﺳﺎﻧﯽ ﺑﺮﺧﯽ از ﮐﻤﯿﺖﻫﺎ ﻣﺎﻧﻨﺪ ﺧﻄﺎي ﺣﺴﯽ، ﺟﺎﺑﻪﺟﺎﯾﯽ و ﺑﺎر ﻣﺤﺎﺳﺒﺎﺗﯽ ﺑﻪﮐﺎر ﮔﺮﻓﺘﻪ ﺷﺪه اﺳﺖ. روش ﭘﯿﺸﻨﻬﺎدي ﻣﻌﺎﯾﺐ روش MPC-MCAﻣﺎﻧﻨﺪ ﺗﺨﻤﯿﻦ ﺗﺠﺮﺑﯽ وﻗﺖﮔﯿﺮ از ﻃﺮﯾﻖ ﺗﮑﺮار آزﻣﻮن و ﺧﻄﺎ ﺑﺮاي ﺗﻌﯿﯿﻦ اﻓﻖﻫﺎي ﮐﻨﺘﺮل و ﭘﯿﺶﺑﯿﻨﯽ را ﺑﺮﻃﺮف ﻣﯽﮐﻨﺪ و در ﻋﯿﻦ ﺣﺎل ﻫﺰﯾﻨﻪ و ﺑﺎر ﻣﺤﺎﺳﺒﺎﺗﯽ را ﺑﻪ ﺣﺪاﻗﻞ ﻣﯽرﺳﺎﻧﺪ. ﻧﺘﺎﯾﺞ ﺷﺒﯿﻪ ﺳﺎزي ﮐﺎرآﻣﺪي روش ﭘﯿﺸﻨﻬﺎدي را ﺑﺮ ﻣﺒﻨﺎي ﺑﻬﺒﻮد ﺧﺮوﺟﯽ ﻋﻤﻠﮑﺮد و ﺑﺎر ﻣﺤﺎﺳﺒﺎﺗﯽ ﻧﺸﺎن ﻣﯽدﻫﺪ.

Marine simulators are effective tools for making a ship feel like driving by creating a similar environment using motion commands. The main problem with simulators is the limited workspace which does not allow them to generate accurate real-time floating movements, so they require a motion synchronization algorithm. Recently, the use of predictive control has become popular in marine simulators. Values of control horizon and future forecast affect the computational load. However, because the designer manually selects these horizons, they are lower than the optimal level. In this paper, a new method based on particle swarm algorithm was used to achieve the best control and forecast horizons by minimizing some periods such as sensory error, displacement and computational load. The proposed method eliminates the disadvantages of the MPC-MCA method such as time-consuming empirical estimation through trial and error for initial control and forecast horizons, while minimizing optimal cost performance and computational load. The simulation results showed the efficiency of the proposed method based on the improvement of performance output and computational load.