No abstract

This study focuses on the development of a supervisory control scheme for power management and operation of an isolated hybrid AC/DC micro-grid, which consists of an AC micro-grid and a DC micro-grid. In the proposed hybrid micro-grid, wind and diesel generators and AC loads are connected to the AC micro-grid, whereas photovoltaic array and DC loads are tied to the DC micro-grid. Moreover, the authors consider two independent battery banks in the AC and DC micro-grids. Furthermore, the AC and the DC micro-grids are coupled through a bidirectional converter, which can act as an inverter or rectifier. The objectives of the proposed supervisory controller are listed as follows: (i) maximum utilisation of renewable energy sources along with satisfying the load power demand in both AC and DC micro-grids, (ii) maintaining state of charge (SOC) of battery banks in both AC and DC micro-grids and (iii) managing the power exchange between the AC and the DC micro-grids while the reliability of the whole system is taken into account. The supervisory controller is formalised using a state machine approach. For these purposes, 15 distinct operation modes are considered. Furthermore, in order to extend the battery life cycle, a fuzzy controller manages the desired SOC controlling the charge and discharge currents. The effectiveness of the proposed supervisory controller is evaluated through extensive numerical simulations.

No abstract

Simultaneous Localization And Mapping (SLAM) is a fundamental problem in mobile robotics. While sparse point-based SLAM methods provide accurate camera localization, the generated maps lack semantic information. On the other hand, state of the art object detection methods provide rich information about entities present in the scene from a single image. This work incorporates a real-time deeplearned object detector to the monocular SLAM framework for representing generic objects as quadrics that permit detections to be seamlessly integrated while allowing the real-time performance. Finer reconstruction of an object, learned by a CNN network, is also incorporated and provides a shape prior for the quadric leading further refinement. To capture the dominant structure of the scene, additional planar landmarks are detected by a CNN-based plane detector and modeled as independent landmarks in the map. Extensive experiments support our proposed inclusion of semantic objects and planar structures directly in the bundle-adjustment of SLAM -Semantic SLAM -that enriches the reconstructed map semantically, while significantly improving the camera localization.

Simultaneous Localization And Mapping (SLAM) is a fundamental problem in mobile robotics. While point-based SLAM methods provide accurate camera localization, the generated maps lack semantic information. On the other hand, state of the art object detection methods provide rich information about entities present in the scene from a single image. This work marries the two and proposes a method for representing generic objects as quadrics which allows object detections to be seamlessly integrated in a SLAM framework. For scene coverage, additional dominant planar structures are modeled as infinite planes. Experiments show that the proposed points-planes-quadrics representation can easily incorporate Manhattan and object affordance constraints, greatly improving camera localization and leading to semantically meaningful maps. The performance of our SLAM system is demonstrated in https://youtu.be/dR-rB9keF8M.

No abstract

No abstract

No abstract