Optimal navigation management in natural inland waterways

Abstract

Today, maritime freight transport is one of the most important factors in the development and growth of world economies in an increasingly interconnected and interdependent world. Within the different scenarios, one of the critical environments whose management is particularly delicate and which today continues to pose different challenges to be solved are natural inland waterways, which are natural logistical navigation channels connected to inland ports. Unlike artificial channels, transportation in natural inland waterways presents formidable challenges due to ever-changing environmental conditions and the unpredictability of different natural phenomena, such as the effect of the tide, in addition to the various operational constraints that arise from environmental, economic and safety issues. This doctoral thesis addresses one of the main challenges in this type of canals, which is the optimal and safe management of navigation within the waterways, all this from a practical approach taking into account the industrial framework in which this work takes place. In the first component of this research, we introduce a methodology aimed at finding optimal navigation plans, which are intended to serve as reference guidelines to which vessel pilots can adhere. These plans are designed with the objective of optimising the waiting and sailing times of vessels, which not only leads to a significant improvement in the efficiency of port and logistics operations, but also has a strong economic impact due to the importance of good channel management on the prestige and positioning of the port. It is important to highlight here the great importance of the effect of the tide, which conditions the time windows in which the channel is available due to the dynamic effect it has on the depth of the channel. To address this issue, a methodology for the search of safe crossing windows is proposed in this work, which guarantees the existence of at least one feasible safe trajectory for each vessel included in the plan. To facilitate the practical implementation of the proposed scheduling methodology, an open-source software tool has been developed. This tool equips navigation planners with a practical solution for optimizing vessel routes by considering real-time data and environmental conditions. The second facet of this thesis tackles the inherent uncertainty associated with inland waterways. Unforeseen incidents, including delays and mechanical failures, can disrupt even the best-laid plans. To counteract these uncertainties, we propose strategies for dynamic rescheduling that allow for real-time optimal adjustments in response to unexpected events. These strategies prioritize safety and aim to minimize the overall impact of the incident. To carry out all this, it is necessary to establish an architecture that allows real-time detection of the incident based on the localization data provided by the vessel, as well as a filter that allows the correct identification of the type of incident that has occurred in order to be able to take the correct action. In the last stage of the present work, we go one step further by looking into the future and propose an autonomous strategy for the control of the vessels, thus making each of the them an autonomous robot capable of navigating the estuary without any human intervention and thus reducing the risk of accidents due to human error, and the need for expert pilots. To this end, the use of distributionally robust control strategies applied to the problem of navigation control in natural inland waterways is presented. The main advantage of this type of methodology is that, unlike conventional robust control strategies that require a perfect identification of the different sources of uncertainty affecting navigation, particularly the effect of the tide, the proposed control strategy is able to improve performance relying only on the knowledge of a very limited set of historical depth data.