Path planning of Antarctic krill targeted fishing net 

based on analytic hierarchy process

Abstract: The exploitation and utilization of Antarctic krill originated in the 1970s. The former Soviet Union adopted the traditional trawl technology for the first trial fishing, followed by Japan, Norway, Poland and other countries to carry out the commercial fishing of Antarctic krill. At present, the commercial fishing of Antarctic krill is all trawling, and the main two kinds of nets are net trawl and truss trawl. The truss trawl has less influence on the horizontal expansion of the net due to the fixed length of the truss, and the fishing efficiency is higher than that of the net trawl. There are single ship operation and mother ship multiship operation in the towing mode. In obtaining the catch, there are artificial collection of the catch through the net and the way of collecting the catch by pumping. In krill fishing in Japan and South Korea, electrohydraulic automatic control of net hoisting and setting has been realized. In the process of trawling, the balance of the trawl and the shape of the net can be adjusted by controlling the acoustic image of the fishing instrument.

Norway is the country with the highest level of automation and fishing efficiency in Antarctic krill fishing at present. The SAGA Sea vessel built by Norway adopts underwater continuous pumping fishing technology for the characteristics of strong clustering and wide area. The main core equipment of the technology system is the shrimp suction pump, the supporting power pipeline of the pump and the krill transport hose. The shrimp suction pump is connected with the net capsule. In the process of trawling, the catch of krill in the net capsule is continuously transported to the fishing vessel by the suction pump through the transport hose. This method is called continuous fishing because it does not need to interrupt the fishing operation by picking up the catch of the net. Continuous fishing greatly reduces the number of times of raising and releasing nets, avoids the squeeze on catches, and ensures the quality of krill catches . It is one of the most advanced krill fishing methods at present. From the perspective of fishing efficiency and technological development trend, the continuous pumping technology with truss trawl is the mainstream of krill fishing technology development in the Antarctica.

Antarctic krill is a very important strategic resource in China’s marine power strategy and Antarctic krill fishing is an important link in the ocean fishing system. But China still lags behind developed countries in the aspects of domestication of fishing equipment, intellectualization of fishing technology and processing equipment, and innovation and upgrading of equipment.The current fishing mode of krill depends on the image of the fishing instrument observed by the captain of the fishing boat and personal experience to determine the fishing depth. The method of artificial determination of krill water layer has a certain lag in the fishing process. If the krill aggregation layer changes, the depth of the krill aggregation abundance center can not be accurately determined in real time, and the fishing depth of the network port can be controlled by adjusting the length of the trawl to target the water layer with high krill aggregation abundance, which will seriously affect the fishing efficiency of Antarctic krill. Therefore, it is urgent to develop an efficient and intelligent moving path planning algorithm for Antarctic krill fishing stringers. When the Antarctic krill is discovered, the moving route of the truss can be planned timely and efficiently to maximize the economic benefit.

In order to overcome the subjectivity of artificial planning and hysteresis, and improve the efficiency of China’s Antarctic krill fishing and automation level. This paper studied the existing route planning algorithm analysis and comparison of the characteristics and applicable scope of various path planning algorithm， comprehensively summarized the domestic and foreign research results about the optimal path planning. This paper also analyzed the technical problems of the moving path planning of krill fishing stringers in the Antarctica. Based on the sonar data collected by the underwater acoustic instrument EK80, this paper proposed a truss moving path planning algorithm for krill fishing operations in the Antarctica. In terms of the determination of fishing depth, the sliding window statistical algorithm was constructed to analyze the data of the fishing instrument, automatically calculate the depth of the largest point of krill cluster abundance, determine the best fishing depth, get rid of the existing method of relying on manual observation of the fishing instrument image, determine the fishing depth of krill, and improve the timeliness of data.The experimental data were collected by underwater acoustic instrument EK80 scientific echo sounder. The data survey time was from 18∶12∶14 on April 18th, 2016 to 18∶31∶53 on April 18th, 2016. The survey area was 63°16′ 46″ S63°18′ 31″ S, 58 °24′ 26″W58 °30′19″W. In addition to krill cluster signals, the echo image data collected also contained a large number of strong and weak interference signals that were higher than krill echo intensity and lower than or close to krill echo intensity. In this paper, the maximum and minimum values of backscattering intensity (Sv, dB) of krill population product collected without interference signal were used as the upper and lower thresholds for eliminating interference signal. The interference signal point after elimination was assigned a value of -999 dB. Echoview software (V8.0.92) was used to preliminarily analyze the water depth range of krill colony distribution from 15 m to 40 m, and the target intensity of krill population product backscattering intensity data was -70 dB and -80 dB, respectively. Firstly, the backscattering intensity data of krill population was analyzed by using sliding window statistical method. The sliding window statistical method needed to design a rectangular data window, and determine the specific position of krill local density center coordinates in the window through calculation. Rectangular fish finder data window was used to simulate artificial observation images, different artificial observation fish finder images were judged in terms of color depth of krill gathered water layer, there would be a certain subjective error and lag. It was more accurate and faster than manual method. In this method, the window sizes were set to 3, 4, 5, 6 and 7 respectively, and 5 groups of local density center coordinate sequences could be obtained. Secondly, in order to make the network port fishing route trajectory executable and reduce the difficulty of calculation, this paper designed a high realtime continuous planning algorithm using cubic Bspline curve as the path planner. Based on the path planning of cubic Bspline curve, the path cluster was constructed by using local density center coordinates. Thirdly, AHP was used to establish a path evaluation model to evaluate the advantages and disadvantages of the path. Taking economy and control ability as criteria, and taking capture rate, path length, curvature and number of inflection points as subcriteria layer, each index was quantified to construct the optimal path system and obtain the optimal path. Finally, the experimental results of real scenes showed that the capture rate of this algorithm was 94.33%, which was 9.8% higher than the traditional method. After many experiments, the average time of path planning was 2.5 s, which could meet the requirements of realtime krill net preplanning.

Keywords: Antarctic krill; analytic hierarchy process; cubic Bspline curve; route planning