Development of methods for managing a trawl complex using predictive modeling on a neural network

The article considers the problem of automation of trawl fishing management in order to increase its efficiency and reduce energy and economic costs when fishing using artificial intelligence technologies and predictive modeling on a neural network. The tasks of long-term, medium-term "on shore" according to the specified criteria and short-term "on the ship" with the use of echolocation forecasting data a have been set. The structure of the system, methods of filling centralized and local databases of catch statistics, artificial neural network learning, systematization of prediction results, calculation and automatic generation of input parameters, configuration are given. The input parameters of the neural network, set by the user (codes of the fishing area, trawl, fishing object, its size, season, average daily water and air temperature, wind speed, wave strength at the time of the intended fishing, vessel type) have been determined together with the calculated values found according to the mathematical model of the trawl system (trawl opening, trawl depth) and user criteria (fishing site code, time of day, trawl speed); output characteristics (catch value, fuel consumption, financial costs); criteria for selecting and grouping output information. The paper presents samples of archived catch data. Application of a mathematical model for the connection of the power and geometric characteristics of the trawl has been justified. The areas of application of the obtained results have been proposed, such as design, production, operation of real trawling systems and development of virtual and augmented reality software and hardware complexes that will allow determining optimal fishing sites taking into account energy and economic costs, collecting catch statistics, automating its analysis, generating analytical reports.

1. Nedostup A. A., Razhev A. O. Primenenie neyronnoy seti dlya upravleniya tralovym promyslom [Application of a neural network for trawl management]. Vestnik Astrakhanskogo gosudarstvennogo tekhnicheskogo universiteta. Seriya: Rybnoe khozyaystvo, Astrakhan', 2021, no. 1, pp. 31–37.

2. Kuznetsov M. Yu., Kuznetsov Yu. A. Gidroakusticheskie metody i sredstva otsenki zapasov ryb i ikh promysla [Hydroacoustic methods and means of assessing fish stocks and their fishery]. Metody i sredstva promyslovoy biogidroakustiki, 2016, part 2, no. 184 (1), pp. 264–294.

3. Kuznetsov M. Yu., Ubarchuk I. A., Polyanichko V. I., Shevtsov V. I., Syrovatkin E. V. Elektronnyy atlas tipovykh akusticheskikh izobrazheniy promyslovykh vidov ryb dal'nevostochnykh morey Rossii [Electronic atlas of typical acoustic images of commercial fish species of the Far Eastern seas of Russia]. Izvestiya TINRO, 2018, no. 193, pp. 57–67.

4. Nedostup A. A., Al'tshul' B. A., Razhev A. O., Dyatchenko S. V., Bedareva O. M., Bagrova A. A. Matematicheskoe modelirovanie povedencheskikh kharakteristik stai ryb pri oblove raznoglubinnym tralom [Mathematical modeling of behavioral chracteristics of a flock of fish when fishing with a midwater trawl]. Morskie intellektual'nye tekhnologii, 2019, no. 4 (46), pp. 181–185.

5. Razhev A. O., Nedostup A. A., L'vova E. E. Sistema avtomatizirovannogo proektirovaniya orudiy promyshlennogo rybolovstva na primere trala [Computer-aided design system for industrial fishing gears on the example of a trawl]. Patent RF, no. 2021619269, 2021.

6. Schmidhuber J. Deep learning in neural networks: An overview. Neural Networks, 2015, no. 61, pp. 85–117.

7. Karpenko E. A. Rybnoe khozyaystvo: obzor. inform. Kompleksnye sravnitel'nye ispytaniya promyslovykh tralov v rayone TSVA [Fisheries: comprehensive comparative tests of fishing trawls in the CVA area]. Moscow, TSNIITEIRX, 1987, iss. DSP-1, 114 p.

8. Nikolaev V. V. Sovershenstvovanie matematicheskoy modeli protsessa lova ryby raznoglubinnym tralom. Diss. kand. tekh. nauk [Improvement of the mathematical model of the fishing process with a midwater trawl. Dis. cand. techn. sci.]. Kaliningrad, 2004, 189 p.