Low-Signal Navigation for Drones

Localization processing service for autonomous navigation without relying on satellite navigation. The service uses onboard sensors and maps to estimate the position of a vehicle when satellite signals are unavailable. It extracts features from captured images and compares them to map features to find matches and determine location. The service is passive, doesn't require network access, and can work in restricted environments.

Source

Controlling a vehicle like a drone or robot in an environment using a probabilistic map of the magnetic field. The probabilistic map relates magnetic field measurements to location probabilities. The vehicle measures magnetic field at its current location and submits that along with timestamp to the probabilistic map to estimate location probability. This probability is used in a stochastic control algorithm to safely navigate the vehicle. The probabilistic mapping allows accurate localization despite magnetic field noise and uncertainty.

Source

Autonomous drone imaging and sensing in datacenters that can continue data collection when network connectivity is lost. The drone uses visual tags placed inside the datacenter to navigate autonomously instead of relying solely on external networks like GPS or Wi-Fi. If connectivity is lost during a mission, the drone switches to autonomous mode using the tags to continue collecting data. This allows uninterrupted drone operations in datacenters with poor network coverage or signal blockage.

Source

Fully autonomous drone flights where the drone takes off, flies according to a predefined plan, and lands without any further user input beyond an initial command and an endpoint target. The drone uses onboard sensors and mapping to navigate and avoid obstacles. It can also communicate with external systems to gather real-time data for planning purposes. The goal is to enable autonomous drone operations in areas with limited connectivity and for situations where remote control is impractical or unsafe.

Source

Improving accuracy of mobile delivery devices like drones and autonomous vehicles indoors or in areas with poor GPS signal by switching to an inertial navigation system like gyroscopes when GPS is less accurate. The system senses GPS accuracy indicators and compares them to thresholds. When GPS is determined to be less accurate, it transitions to using the inertial navigation system for positioning. This allows more accurate tracking indoors and in GPS-challenged areas.

Source

Reliable positioning for drones in areas with poor GPS reception by using a network of ground base stations. When a drone parks, it communicates with nearby base stations to obtain their coordinates. It then estimates the coordinate of a farther base station it can't directly see. This allows the drone to use the nearby base stations as a network to navigate when GPS is unavailable. The parking station stores the base station coordinates for the drone to use during flight.

Source

Ensuring reliable communication in remote or disaster-affected areas is a technical challenge due to unplanned deployment and mobilization, meaning placement difficulties high operation costs of conventional telecommunications infrastructures. To address this problem, unmanned aerial vehicles (UAVs) have emerged as an excellent alternative provide quick connectivity regions at reasonable cost. However, the limited battery autonomy UAVs restricts their flight service time. This paper proposes system based on tethered UAV (T-UAV) capable continuous through wired power network connected ground station. The communications low-cost devices, such Raspberry Pi platforms, offers wireless IP telephony services, providing high-quality communication. Experimental tests assessed consumption, stability, data transmission performance. Our results prove that T-UAV, quadcopter drone, operates stably 16 V 20 A, ensuring consistent VoIP height 10 m with low latency. These experimental findings underscore potential T-UAVs cost-effective alternatives for extending networks regions, emergency scenarios, ... Read More

Source

System for geophysical fields sensing-based navigation in environments with intermittent or no data connectivity, where GPS signals may be unavailable. The system uses an offline baseline estimation model stored on the navigation object to estimate geophysical fields like magnetic fields when online model data is not available. It selectively uses the offline model or online data to estimate fields at specified times based on control logic. This allows reliable navigation without constant connectivity.

Source

Urban Air Mobility (UAM) requires reliable communication and surveillance infrastructures to ensure safe Unmanned Aerial Vehicle (UAV) operations in dense metropolitan environments. However, urban infrastructure is inherently heterogeneous, leading significant spatial variations monitoring performance. This study proposes a unified framework that integrates readiness assessment with Deep Reinforcement Learning (DRL)-based UAV path planning. Using Singapore as representative case, we employ data-driven methodology combining clustering analysis situ measurements estimate the citywide distribution of quality. We then introduce an infrastructure-aware planning algorithm based on Double Q-Network (DQN) convolutional architecture, which enables UAVs learn efficient trajectories while avoiding blind zones. Extensive simulations demonstrate proposed approach significantly improves success rates, reduces traversal through poorly monitored regions, maintains high navigation efficiency. These results highlight potential modeling DRL support performance-aware airspace inform future UAM governanc... Read More

Source

Abstract Existing collective navigation systems for drone swarms typically rely on the communication between drones, which limits application in specific mission scenarios and reduces robustness against interference. To address this challenge, a communication-free method enhanced by relative state estimation is proposed study. It consists of three key components: visual perception localization, estimation, swarm motion decision. First, sensors are employed to detect nearby drones real time calculate their positions. Second, an optimized model set adaptive interacting multiple (OMSA-IMM) filtering algorithm fuse predicted states from with measurements achieve continuous high-precision positioning. The fused then fed back into decision high-level control. Finally, effectiveness validated through numerical simulations real-world flight experiments. results demonstrate that significantly enhances accuracy localization improves performance algorithm, enabling cohesive collision-free communication-denied environments.

Source

In Japan, natural disasters occur frequently. Serious may cause damage to traffic networks and telecommunication infrastructures, leading the occurrence of isolated disaster areas. this article, unmanned aerial vehicles (UAVs) are used for data collection instead unavailable ground-based stations in Detailed information about situation will be collected from user equipment (UE) by a UAV through flyhoverfly procedure, then sent response headquarters relief. However, mission completion time minimization becomes crucial task, considering requirement rapid battery constraint UAVs. Therefore, author proposed three-dimensional flight trajectory, discussing optimal altitude placement hovering points transforming original problem K-means clustering into location set cover (LSCP) that can solved via genetic algorithm (GA) approach. The simulation results have shown feasibility method reduce time.

Source

Positioning information processing method for vehicles that improves accuracy when GPS is unavailable. It combines dead reckoning using inertial sensors with visual odometry using cameras. The method involves obtaining positioning from dead reckoning, then correcting it with visual odometry. Historical dead reckoning data is used to determine distances between past and current positions. These distances are used to correct the current visual odometry. The corrected visual odometry is then used as the final position. This iterative correction and compensation improves positioning accuracy when GPS is unavailable.

Source

Robotic drone system for autonomous inventory management in indoor environments like warehouses. The drone uses onboard cameras, sensors, and processing to generate a 3D map of the warehouse, estimate its own position and orientation in the map, and autonomously navigate through the map while avoiding obstacles. This allows the drone to operate without GPS and external monitoring in environments where GPS signals are weak or unavailable. The drone can also find optimal paths through the map using its own generated 3D data.

Source

This article considers a multi-cell wireless network comprising of conventional user equipment (UE), sensor devices and unmanned aerial vehicles (UAVs) or drones. UAVs are used to assist base station, e.g., improve coverage collect data from devices. The problem at hand is optimize the (i) sub-carrier assigned cell (ii) position each UAV, (iii) transmit power following nodes: stations UAVs. We outline two-stage approach maximize fairness-aware sum-rate UE In first stage, genetic algorithm (GA)-based assign sub-band all cells determine location UAV. Then, in second linear program results demonstrate that our proposed achieves approximately 97.43% optimal obtained via brute-force search. It also attains on average 98.78% performance computationally intensive benchmark requires over 478% longer run-time. Furthermore, it outperforms GA-based allocation heuristic by 221.39%.

Source

Enhanced self-localization for drones using reconfigurable intelligent surfaces (RIS) and fusion algorithm integration to improve accuracy in urban environments where GPS signals are prone to interference. The system fuses RIS-based positioning signals with other sources like GPS, cellular, and IMU to determine drone location using a data fusion algorithm like an extended Kalman filter. Strategically placed RIS antennas provide signals to aid drone navigation along paths and landing at destinations.

Source

Unmanned aerial vehicle (UAV) with collision avoidance using onboard sensors. The UAV has multiple optical radars to measure distances to obstacles, as well as a visual sensing module for SLAM (simultaneous localization and mapping) to determine obstacle positions. An edge processor converts the SLAM data to match the radar coordinates. The UAV's flight controller uses all the sensor data to maneuver around obstacles. This allows the UAV to effectively avoid collisions without relying solely on GPS or external positioning.

Source

Abstract To improve the accuracy and efficiency of autonomous navigation for indoor UAVs (Unmanned Aerial Vehicle) in complex substation environments, this paper proposes a multi-sensor fusion-based method. An UAV platform integrating multi-dimensional LiDAR, depth cameras, edge computing devices was designed built. The method uses AprilTag markers as visual references constructs maps based on laser SLAM(Simultaneous Localization And Mapping) technology. A relocation scheme that incorporates information is proposed, Dijkstra DWA (Dynamic Window Approach) algorithms are used to plan global local paths UAV, respectively. In GPS-denied environment, system achieves precise localization environmental point cloud mapping, ensuring accurate positioning path planning. Multi-scenario experiments ROS (Robot Operating System) environment demonstrate provides high strong robustness.

Source

Structured light navigation aid for autonomous vehicles that uses emitted structured light patterns to improve navigation in GNSS-denied environments. The aid has a movable structured light emitter that projects patterns onto surfaces. A receiver captures reflected light to calculate navigation info. This provides a fixed, drift-free source vs. GNSS-prone drift. The emitter can change direction and pattern. This allows scanning to find suitable landing spots based on features.

Source

Unmanned aerial vehicle (UAV) localization system that enables precise determination of its absolute position in complex environments through machine learning-based 3D point cloud registration. The system employs a trained model that generates both semantic and depth representations of the environment from camera images, which are then registered against a pre-existing point cloud. This registration process enables accurate determination of the UAV's position and orientation, even in environments with GPS signal loss or sensor malfunctions. The system can be used for autonomous missions where GPS is unavailable or unreliable.

Source

Optimizing UAV flight paths through AI-driven radio condition adaptation. The method enables UAVs to automatically adjust their flight paths in response to changing radio conditions, leveraging machine learning models to determine optimal path modifications. The approach integrates with UAV networks to dynamically configure network nodes and UAVs based on real-time radio environment conditions, ensuring seamless path adaptation while maintaining network stability.

Source