15 March 2021 in Nature Machine Intelligence
1 December 2021 in Evolutionary Computation
Robots operating in the real world will experience a range of different environments and tasks. It is essential for the robot to have the ability to adapt to its surroundings to work efficiently in changing conditions. Evolutionary robotics aims to solve this by optimizing both the control and body (morphology) of a robot, allowing adaptation to internal, as well as external factors. Most work in this field has been done in physics simulators, which are relatively simple and not able to replicate the richness of interactions found in the real world. Solutions that rely on the complex interplay among control, body, and environment are therefore rarely found. In this article, we rely solely on real-world evaluations and apply evolutionary search to yield combinations of morphology and control for our mechanically self-reconfiguring quadruped robot. We evolve solutions on two distinct physical surfaces and analyze the results in terms of both control and morphology. We then transition to two previously unseen surfaces to demonstrate the generality of our method. We find that the evolutionary search finds high-performing and diverse morphology-controller configurations by adapting both control and body to the different properties of the physical environments. We additionally find that morphology and control vary with statistical significance between the environments. Moreover, we observe that our method allows for morphology and control parameters to transfer to previously unseen terrains, demonstrating the generality of our approach.
22 November 2019 in Communications in Computer and Information Science
23 July 2025 in IEEE Journal of Photovoltaics
As the installed capacity of photovoltaic power plants continues its near exponential growth, cost-efficient operation and maintenance strategies become increasingly crucial. Aerial infrared thermography has enabled fast and robust fault detection in utility-scale PV plants. In this article, we explore two key approaches to improve inspection efficiency: increase the flight altitude and deploy swarms of unmanned aerial vehicles. A larger imaging distance expands the field of view but reduces fault detectability and georeferencing accuracy. In this work, we study the tradeoff between inspection efficiency and data quality for automatic fault detection and localization. The YOLO11 machine learning model was trained to detect defects in thermal images, and its performance was evaluated to vary imaging distances and camera pitch angles. Fault detection remained robust up to approximately 80 m, but georeferencing error became the primary limiting factor. Finally, we conduct a UAV swarm-based inspection of a PV plant, integrating automatic fault detection and localization, and compare the results with ground truth data.