Loading...
5 results
Search Results
Now showing 1 - 5 of 5
- Communication Protocol for Unmanned Vehicles : An Architectural ApproachPublication . Ramos, João; Ribeiro, Roberto; Safadinho, David; Barroso, João; Pereira, AntonioThe great potential of Unmanned Vehicles (UV) for services is supported by the evolution of drones and their flexible ability to help us performing dangerous, boring and difficult tasks. The common communication solutions e.g., radio, consider a pilot and a vehicle in the same location. Besides, the alternatives based on Ground Control Stations (GCS) are difficult to configure, which makes it difficult for a common user and a developer to use and/or implement new services for drones. This work proposes a solution with a new messaging protocol to overcome the previous problems, allowing controlling a UV in a remote location after a simple configuration. The implementation of this approach is based on a cloud platform, responsible for UV and user management, and on TCP/IP WebSockets for the user to control remotely located vehicles, anytime and anywhere. This research starts with the analysis of the prior work regarding UV communication technologies and the available GCS. Minding the identified problems, we designed an architecture that represents a cloud platform and multiple users and UVs in different networks. To prototype this architecture, we developed the user interface, the platform and the UV control module. The functional assessment considers a generic simplified and controlled scenario, focusing on the real-time control and video stream. The results confirmed the possibility to control a UV in a different location, in real-time, with an average response time of 25ms. This work provides valuable insights regarding the communication in the fields of anytime and anywhere device control and vehicle-based services.
- Applying deep learning to real-time UAV-based forest monitoring: Leveraging multi-sensor imagery for improved resultsPublication . Marques, Tomás; Carreira, Samuel; Miragaia, Rolando; Ramos, João; Pereira, AntónioRising global fire incidents necessitate effective solutions, with forest surveillance emerging as a crucial strategy. This paper proposes a complete solution using technology that integrates visible and infrared spectrum images through Unmanned Aerial Vehicles (UAVs) for enhanced detection of people and vehicles in forest environments. Unlike existing computer vision models relying on single-sensor imagery, this approach overcomes limitations posed by limited spectrum coverage, particularly addressing challenges in low-light conditions, fog, or smoke. The developed 4-channel model uses both types of images to take advantage of the strengths of each one simultaneously. This article presents the development and implementation of a solution for forest monitoring ranging from the transmission of images captured by a UAV to their analysis with an object detection model without human intervention. This model consists of a new version of the YOLOv5 (You Only Look Once) architecture. After the model analyzes the images, the results can be observed on a web platform on any device, anywhere in the world. For the model training, a dataset with thermal and visible images from the aerial perspective was captured with a UAV. From the development of this proposal, a new 4- channel model was created, presenting a substantial increase in precision and mAP (Mean Average Precision) metrics compared to traditional SOTA (state-of-the-art) models that only make use of red, green, and blue (RGB) images. Allied with the increase in precision, we confirmed the hypothesis that our model would perform better in conditions unfavorable to RGB images, identifying objects in situations with low light and reduced visibility with partial occlusions. With the model’s training using our dataset, we observed a significant increase in the model’s performance for images in the aerial perspective. This study introduces a modular system architecture featuring key modules: multisensor image capture, transmission, processing, analysis, and results presentation. Powered by an innovative object detection deep-learning model, these components collaborate to enable real-time, efficient, and distributed forest monitoring across diverse environments.
- UAV Landing Using Computer Vision Techniques for Human DetectionPublication . Safadinho, David; Ramos, João; Ribeiro, Roberto; Filipe, Vítor; Barroso, João; Pereira, AntónioThe capability of drones to perform autonomous missions has led retail companies to use them for deliveries, saving time and human resources. In these services, the delivery depends on the Global Positioning System (GPS) to define an approximate landing point. However, the landscape can interfere with the satellite signal (e.g., tall buildings), reducing the accuracy of this approach. Changes in the environment can also invalidate the security of a previously defined landing site (e.g., irregular terrain, swimming pool). Therefore, the main goal of this work is to improve the process of goods delivery using drones, focusing on the detection of the potential receiver. We developed a solution that has been improved along its iterative assessment composed of five test scenarios. The built prototype complements the GPS through Computer Vision (CV) algorithms, based on Convolutional Neural Networks (CNN), running in a Raspberry Pi 3 with a Pi NoIR Camera (i.e., No InfraRed-without infrared filter). The experiments were performed with the models Single Shot Detector (SSD) MobileNet-V2, and SSDLite-MobileNet-V2. The best results were obtained in the afternoon, with the SSDLite architecture, for distances and heights between 2.5-10 m, with recalls from 59%-76%. The results confirm that a low computing power and cost-effective system can perform aerial human detection, estimating the landing position without an additional visual marker.
- Service Oriented Platform for Drones CompetitionPublication . Ramos, João; Safadinho, David; Ribeiro, Roberto; Pereira, António Manuel de JesusDrones are facing a huge success. They are available almost anywhere, either online or at a near shop, not only for enthusiasts but also for professionals of many areas, such as agriculture, construction, multimedia and military missions. In the same way, online platforms are growing fast. Many applications that we were used to be installed and ran locally in our devices are now deployed in the internet, with the benefits like high availability, high scalability and high compatibility with almost any laptop, computer, smartphone or tablet. Thus, the high level of interest and utility of drones and the crescent demand of service-oriented platforms surged up to be a great fusion of concepts. Nowadays, training and competition requires users to have their own aircraft(s) and to move, usually, to an arena where they can play with each other. Having it in consideration, in this work it's proposed a platform that combines real-time online gaming and drones' training and competition, in which users from around the world can socialize and play together. This paper runs through an overview of the concept, presenting the specifications and the requirements that the platform must meet to accomplish the desired objective. The proposed platform provides access to anyone, anytime, anywhere and with any device. It opens the possibility for users to communicate with each other and to control aircrafts with real-time video transmission, only achievable due to the distributed architecture designed, that minimizes the existence of single points of failure.
- UAV for Everyone : An Intuitive Control Alternative for Drone Racing CompetitionsPublication . Ribeiro, Roberto; Ramos, João; Safadinho, David; Pereira, António Manuel de JesusAs many technological innovations, the Unmanned Aerial Vehicles (UAV) were created for military purposes. However, these aircrafts are now a reality accessible for common civilians and are used to explore new possibilities in areas like cinema, photography or surveillance. The technology easily reached other application fields and turned into a new sports category as well, supported by many drone racing competitions that happen around the world. There are many alternatives to control a quadcopter, but the most accurate relies on a robust and heavy handheld remote controller that requires dexterity to be a good pilot. This limits the control of the UAV to users comfortable with this kind of interaction, discarding people with reduced fine motor skills, hand malformations or elders with low learning capacity from the possibility of piloting a quadcopter and participate in competition events. An alternative to these remote controllers should be developed to overpass the mentioned disadvantages. The solution proposed in this paper is an intuitive and accurate control system made of two small and lightweight wearable devices capable of detecting patterns based on motion events. These patterns are then sent to a mobile application responsible for controlling the drone. Wearing the described devices like a high-tech glove, the users can pilot the UAV through simple and intuitive upper limb movements, being the changes in the orientation of each device over time identified as a specific pattern. This solution presents a new approach to the control of UAV that improves the ease of piloting them and decreases the time that is required to learn how to handle them.