This paper presents a proof-of-concept optical detector array sensor system to be used in Unmanned Underwater Vehicle (UUV) navigation. The performance of the developed optical detector array was evaluated for its capability to estimate the position, orientation and forward velocity of UUVs with respect to a light source fixed in underwater. The evaluations were conducted through Monte Carlo simulations and empirical tests under a variety of motion configurations. Monte Carlo simulations also evaluated the system total propagated uncertainty (TPU) by taking into account variations in the water column turbidity, temperature and hardware noise that may degrade the system performance. Empirical tests were conducted to estimate UUV position and velocity during its navigation to a light beacon. Monte Carlo simulation and empirical results support the use of the detector array system for optics based position feedback for UUV positioning applications.

Trust and confidence in autonomous behavior is required to send autonomous vehicles into operational missions. The authors introduce the Performance Evaluation and Review Framework Of Robotic Missions (PERFORM), a framework to enable a rigorous and replicable autonomy test environment, thereby filling the void between that of merely simulating autonomy and that of completing true field missions. A generic architecture for defining the missions under test is proposed and a unique Interval Type-2 Fuzzy Logic approach is used as the foundation for the mathematically rigorous autonomy evaluation framework. The test environment is designed to aid in (1) new technology development (i.e., providing direct comparisons and quantitative evaluations between autonomy algorithms), (2) the validation of the performance of specific autonomous platforms, and (3) the selection of the appropriate robotic platform(s) for a given mission type (e.g., for surveying, surveillance, search and rescue). Three case studies are presented to apply the metric to various test scenarios. Results demonstrate the flexibility of the technique with the ability to tailor tests to the user’s design requirements accounting for different priorities related to acceptable risks and goals of a given mission.

This paper explores the emerging synergy between control theory and machine learning in robotics, focusing on methods that combine model-based strategies with data-driven adaptation. The authors highlight how classical techniques, such as model predictive control and adaptive control, are being enhanced by reinforcement learning, imitation learning, and neural models to address challenges in complex, uncertain environments. Emphasis is placed on real-world platforms (e.g., legged systems, aerial robots, and manipulators) with special attention to advanced domains such as multi-agent systems and coordination. The authors, in addition, establish a taxonomy to categorize these hybrid approaches as “learning-for-control”, “control-for-learning”, or “co-designed architectures”. This paper also reflects upon key open problems, including sim-to-real transfer, safety, and the need for verifiable learning-based controllers, all facets that help to outline a roadmap for future research.