Navigation Architecture for a Catamaran-Based Robot: Vision-Focused Obstacle Avoidance with Complementary Sonar Sensing in Aquaculture Monitoring

Figueroa J.A.R.; Reategui R.A.H.; Flores P.A.F.; Cuentas C.V.; Delgado G.J.C.; Cayo E.H. · Proceedings - 2025 2nd International Conference on Electronic Circuits and Signaling Technologies, ICECST 2025 (2025) · DOI: 10.1109/icecst66106.2025.11307629

Resumen

This paper presents the design, implementation, and validation of an multi-sensor navigation system for au- tonomous environmental monitoring in aquatic environments. The system is built on a custom catamaran platform equipped with a multi-sensor perception suite combining stereo vision (ZED 2i) and scanning sonar (PING360) to detect obstacles above and below the water surface. Development followed a structured methodology including ROS/Gazebo simulation with sensor emulation, algorithm integration, and real-world testing. A reactive Vector Field Histogram (VFH) algorithm ensures safe and efficient navigation amidst static and dynamic obstacles. Field trials at San Jose' de Uzuña Reservoir (Arequipa, Peru) demonstrated robust real-time obstacle avoidance by the vision system, complemented by sonar-based detection of submerged hazards. The platform supports onboard processing and modular deployment with minimal human intervention. This integrated solution significantly improves autonomous navigation reliability and data collection efficiency in complex aquatic environments, with applications in aquaculture monitoring, water quality assessment, and environmental surveying.

Instituciones: Universidad Católica San Pablo