Autonomous drone hunter operating by deep learning and all-onboard computations in GPS-denied environments.
Ontology highlight
ABSTRACT:
SUBMITTER: Wyder PM
PROVIDER: S-EPMC6860441 | biostudies-literature | 2019
REPOSITORIES: biostudies-literature
Publications
Autonomous drone hunter operating by deep learning and all-onboard computations in GPS-denied environments.
Wyder Philippe Martin PM Chen Yan-Song YS Lasrado Adrian J AJ Pelles Rafael J RJ Kwiatkowski Robert R Comas Edith O A EOA Kennedy Richard R Mangla Arjun A Huang Zixi Z Hu Xiaotian X Xiong Zhiyao Z Aharoni Tomer T Chuang Tzu-Chan TC Lipson Hod H
PloS one 20191118 11
This paper proposes a UAV platform that autonomously detects, hunts, and takes down other small UAVs in GPS-denied environments. The platform detects, tracks, and follows another drone within its sensor range using a pre-trained machine learning model. We collect and generate a 58,647-image dataset and use it to train a Tiny YOLO detection algorithm. This algorithm combined with a simple visual-servoing approach was validated on a physical platform. Our platform was able to successfully track an ...[more]
Similar Datasets
Project description:In recent years, unmanned aerial vehicles (UAVs) have been applied in many fields owing to their mature flight control technology and easy-to-operate characteristics. No doubt, these UAV-related applications rely heavily on location information provided by the positioning system. Most UAVs nowadays use a global navigation satellite system (GNSS) to obtain location information. However, this outside-in 3rd party positioning system is particularly susceptible to environmental interference and cannot be used in indoor environments, which limits the application diversity of UAVs. To deal with this problem, in this paper, a stereo-based visual simultaneous localization and mapping technology (vSLAM) is applied. The presented vSLAM algorithm fuses onboard inertial measurement unit (IMU) information to further solve the navigation problem in an unknown environment without the use of a GNSS signal and provides reliable localization information. The overall visual positioning system is based on the stereo parallel tracking and mapping architecture (S-PTAM). However, experiments found that the feature-matching threshold has a significant impact on positioning accuracy. Selection of the threshold is based on the Hamming distance without any physical meaning, which makes the threshold quite difficult to set manually. Therefore, this work develops an online adaptive matching threshold according to the keyframe poses. Experiments show that the developed adaptive matching threshold improves positioning accuracy. Since the attitude calculation of the IMU is carried out based on the Mahony complementary filter, the difference between the measured acceleration and the gravity is used as the metric to online tune the gain value dynamically, which can improve the accuracy of attitude estimation under aggressive motions. Moreover, a static state detection algorithm based on the moving window method and measured acceleration is proposed as well to accurately calculate the conversion mechanism between the vSLAM system and the IMU information; this initialization mechanism can help IMU provide a better initial guess for the bundle adjustment algorithm (BA) in the tracking thread. Finally, a performance evaluation of the proposed algorithm is conducted by the popular EuRoC dataset. All the experimental results show that the developed online adaptive parameter tuning algorithm can effectively improve the vSLAM accuracy and robustness.
| S-EPMC9610021 | biostudies-literature
Project description:This paper proposes a novel infrared camera array guidance system with capability to track and provide real time position and speed of a fixed-wing Unmanned air vehicle (UAV) during a landing process. The system mainly include three novel parts: (1) Infrared camera array and near infrared laser lamp based cooperative long range optical imaging module; (2) Large scale outdoor camera array calibration module; and (3) Laser marker detection and 3D tracking module. Extensive automatic landing experiments with fixed-wing flight demonstrate that our infrared camera array system has the unique ability to guide the UAV landing safely and accurately in real time. Moreover, the measurement and control distance of our system is more than 1000 m. The experimental results also demonstrate that our system can be used for UAV automatic accurate landing in Global Position System (GPS)-denied environments.
| S-EPMC5038671 | biostudies-literature
Project description:In V1, local circuitry depends on the position in the orientation map: close to pinwheel centers, recurrent inputs show variable orientation preferences; within iso-orientation domains, inputs are relatively uniformly tuned. Physiological properties such as cell's membrane potentials, spike outputs, and temporal characteristics change systematically with map location. We investigate in a firing rate and a Hodgkin-Huxley network model what constraints these tuning characteristics of V1 neurons impose on the cortical operating regime. Systematically varying the strength of both recurrent excitation and inhibition, we test a wide range of model classes and find the likely models to account for the experimental observations. We show that recent intracellular and extracellular recordings from cat V1 provide the strongest evidence for a regime where excitatory and inhibitory recurrent inputs are balanced and dominate the feed-forward input. Our results are robust against changes in model assumptions such as spatial extent and strength of lateral inhibition. Intriguingly, the most likely recurrent regime is in a region of parameter space where small changes have large effects on the network dynamics, and it is close to a regime of "runaway excitation," where the network shows strong self-sustained activity. This could make the cortical response particularly sensitive to modulation.
| S-EPMC2722429 | biostudies-literature
Project description:The objective of autonomous robotic additive manufacturing for construction in the architectural scale is currently being investigated in parts both within the research communities of computational design and robotic fabrication (CDRF) and deep reinforcement learning (DRL) in robotics. The presented study summarizes the relevant state of the art in both research areas and lays out how their respective accomplishments can be combined to achieve higher degrees of autonomy in robotic construction within the Architecture, Engineering and Construction (AEC) industry. A distributed control and communication infrastructure for agent training and task execution is presented, that leverages the potentials of combining tools, standards and algorithms of both fields. It is geared towards industrial CDRF applications. Using this framework, a robotic agent is trained to autonomously plan and build structures using two model-free DRL algorithms (TD3, SAC) in two case studies: robotic block stacking and sensor-adaptive 3D printing. The first case study serves to demonstrate the general applicability of computational design environments for DRL training and the comparative learning success of the utilized algorithms. Case study two highlights the benefit of our setup in terms of tool path planning, geometric state reconstruction, the incorporation of fabrication constraints and action evaluation as part of the training and execution process through parametric modeling routines. The study benefits from highly efficient geometry compression based on convolutional autoencoders (CAE) and signed distance fields (SDF), real-time physics simulation in CAD, industry-grade hardware control and distinct action complementation through geometric scripting. Most of the developed code is provided open source. Supplementary Information The online version contains supplementary material available at 10.1007/s41693-022-00069-0.
| S-EPMC9125977 | biostudies-literature
Project description:First-person view (FPV) drone racing is a televised sport in which professional competitors pilot high-speed aircraft through a 3D circuit. Each pilot sees the environment from the perspective of their drone by means of video streamed from an onboard camera. Reaching the level of professional pilots with an autonomous drone is challenging because the robot needs to fly at its physical limits while estimating its speed and location in the circuit exclusively from onboard sensors1. Here we introduce Swift, an autonomous system that can race physical vehicles at the level of the human world champions. The system combines deep reinforcement learning (RL) in simulation with data collected in the physical world. Swift competed against three human champions, including the world champions of two international leagues, in real-world head-to-head races. Swift won several races against each of the human champions and demonstrated the fastest recorded race time. This work represents a milestone for mobile robotics and machine intelligence2, which may inspire the deployment of hybrid learning-based solutions in other physical systems.
| S-EPMC10468397 | biostudies-literature
Project description:Nowadays, more than half of the world's population lives in urban areas, and this number continues increasing. Consequently, there are more and more scientific publications that analyze health problems of people associated with living in these highly urbanized locations. In particular, some of the recent work has focused on relating people's health to the quality and quantity of urban green areas. In this context, and considering the huge amount of land area in large cities that must be supervised, our work seeks to develop a deep learning-based solution capable of determining the level of health of the land and to assess whether it is contaminated. The main purpose is to provide health institutions with software capable of creating updated maps that indicate where these phenomena are presented, as this information could be very useful to guide public health goals in large cities. Our software is released as open source code, and the data used for the experiments presented in this paper are also freely available.
| S-EPMC6928838 | biostudies-literature
Project description:In this paper, an onboard vision-based system for the autonomous landing of a low-cost quadrotor is presented. A novel landing pad with different optical markers sizes is carefully designed to be robustly recognized at different distances. To provide reliable pose information in a GPS (Global Positioning System)-denied environment, a vision algorithm for real-time landing pad recognition and pose estimation is implemented. The dynamic model of the quadrotor is established and a system scheme for autonomous landing control is presented. A series of autonomous flights have been successfully performed, and a video of the experiment is available online. The efficiency and accuracy of the presented vision-based system is demonstrated by using its position and attitude estimates as control inputs for the autonomous landing of a self-customized quadrotor.
| S-EPMC6874798 | biostudies-literature
Project description:Purpose: Most built environment studies have quantified characteristics of the areas around participants' homes. However, the environmental exposures for physical activity (PA) are spatially dynamic rather than static. Thus, merged accelerometer and global positioning system (GPS) data were utilized to estimate associations between the built environment and PA among adults. Methods: Participants (N = 142) were recruited on trails in Massachusetts and wore an accelerometer and GPS unit for 1-4 days. Two binary outcomes were created: moderate-to-vigorous PA (MVPA vs. light PA-to-sedentary); and light-to-vigorous PA (LVPA vs. sedentary). Five built environment variables were created within 50-meter buffers around GPS points: population density, street density, land use mix (LUM), greenness, and walkability index. Generalized linear mixed models were fit to examine associations between environmental variables and both outcomes, adjusting for demographic covariates. Results: Overall, in the fully adjusted models, greenness was positively associated with MVPA and LVPA (odds ratios [ORs] = 1.15, 95% confidence interval [CI] = 1.03, 1.30 and 1.25, 95% CI = 1.12, 1.41, respectively). In contrast, street density and LUM were negatively associated with MVPA (ORs = 0.69, 95% CI = 0.67, 0.71 and 0.87, 95% CI = 0.78, 0.97, respectively) and LVPA (ORs = 0.79, 95% CI = 0.77, 0.81 and 0.81, 95% CI = 0.74, 0.90, respectively). Negative associations of population density and walkability with both outcomes reached statistical significance, yet the effect sizes were small. Conclusions: Concurrent monitoring of activity with accelerometers and GPS units allowed us to investigate relationships between objectively measured built environment around GPS points and minute-by-minute PA. Negative relationships between street density and LUM and PA contrast evidence from most built environment studies in adults. However, direct comparisons should be made with caution since most previous studies have focused on spatially fixed buffers around home locations, rather than the precise locations where PA occurs.
| S-EPMC6701185 | biostudies-literature
Project description:Soft robots that can harvest energy from environmental resources for autonomous locomotion is highly desired; however, few are capable of adaptive navigation without human interventions. Here, we report twisting soft robots with embodied physical intelligence for adaptive, intelligent autonomous locomotion in various unstructured environments, without on-board or external controls and human interventions. The soft robots are constructed of twisted thermal-responsive liquid crystal elastomer ribbons with a straight centerline. They can harvest thermal energy from environments to roll on outdoor hard surfaces and challenging granular substrates without slip, including ascending loose sandy slopes, crossing sand ripples, escaping from burying sand, and crossing rocks with additional camouflaging features. The twisting body provides anchoring functionality by burrowing into loose sand. When encountering obstacles, they can either self-turn or self-snap for obstacle negotiation and avoidance. Theoretical models and finite element simulation reveal that such physical intelligence is achieved by spontaneously snapping-through its soft body upon active and adaptive soft body-obstacle interactions. Utilizing this strategy, they can intelligently escape from confined spaces and maze-like obstacle courses without any human intervention. This work presents a de novo design of embodied physical intelligence by harnessing the twisting geometry and snap-through instability for adaptive soft robot-environment interactions.
| S-EPMC9295757 | biostudies-literature
Project description:Counter-drone technology by using artificial intelligence (AI) is an emerging technology and it is rapidly developing. Considering the recent advances in AI, counter-drone systems with AI can be very accurate and efficient to fight against drones. The time required to engage with the target can be less than other methods based on human intervention, such as bringing down a malicious drone by a machine-gun. Also, AI can identify and classify the target with a high precision in order to prevent a false interdiction with the targeted object. We believe that counter-drone technology with AI will bring important advantages to the threats coming from some drones and will help the skies to become safer and more secure. In this study, a deep reinforcement learning (DRL) architecture is proposed to counter a drone with another drone, the learning drone, which will autonomously avoid all kind of obstacles inside a suburban neighborhood environment. The environment in a simulator that has stationary obstacles such as trees, cables, parked cars, and houses. In addition, another non-malicious third drone, acting as moving obstacle inside the environment was also included. In this way, the learning drone is trained to detect stationary and moving obstacles, and to counter and catch the target drone without crashing with any other obstacle inside the neighborhood. The learning drone has a front camera and it can capture continuously depth images. Every depth image is part of the state used in DRL architecture. There are also scalar state parameters such as velocities, distances to the target, distances to some defined geofences and track, and elevation angles. The state image and scalars are processed by a neural network that joints the two state parts into a unique flow. Moreover, transfer learning is tested by using the weights of the first full-trained model. With transfer learning, one of the best jump-starts achieved higher mean rewards (close to 35 more) at the beginning of training. Transfer learning also shows that the number of crashes during training can be reduced, with a total number of crashed episodes reduced by 65%, when all ground obstacles are included.
| S-EPMC7219666 | biostudies-literature