Model for determining pairwise-adaptive minimal allowable values of CPA and TCPA based on fuzzy logic

The article addresses the problem of determining pairwise-adaptive minimum allowable values of CPA and TCPA based on fuzzy logic. The main idea of the proposed model is that collision-avoidance threshold values are defined not as fixed values, but as functions of the characteristics of an interacting pair of vessels. Within the framework of this study, several pairs of vessels with different deadweights are considered, for each of which safe collision-avoidance conditions must be determined. In contrast to traditional approaches, the principle of pairwise adaptation is introduced, according to which an individual minimum allowable value is determined for each pair of vessels based on their deadweights. A model with two input variables the deadweights of the interacting vessels has been developed, forming functional dependencies of CPAmin and TCPAmin on the input variables . An important stage of the study was the comparison of two rule-base configurations: a full rule base consisting of 16 rules and a simplified rule base consisting of four rules. Analysis of the tables shows that the 4-rule and 16-rule configurations produce fundamentally different distributions of threshold values. In the case of the full rule base, the threshold values range from 1.20 to 2.0 nautical miles for CPA and from 12 to 20 minutes for TCPA, with CPAmin and TCPAmin increasing monotonically as vessel deadweight increases. At the same time, the simplified rule base demonstrates a different distribution pattern: for CPA, most values are concentrated around 1.75 nautical miles, and differences between vessel pairs with regard to their class are not clearly expressed. Thus, the four-rule configuration cannot be considered an example of pairwise adaptation, but rather a quasi-constant distribution of threshold values. It is concluded that the use of pairwise-adaptive thresholds improves the adequacy of collisionrisk assessment and enables better-grounded decision-making. The results of the study are intended to be used in further work to develop control criteria and algorithms for collision avoidance in multi-ship scenarios .