AU2021102531A4 - Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning - Google Patents
Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning Download PDFInfo
- Publication number
- AU2021102531A4 AU2021102531A4 AU2021102531A AU2021102531A AU2021102531A4 AU 2021102531 A4 AU2021102531 A4 AU 2021102531A4 AU 2021102531 A AU2021102531 A AU 2021102531A AU 2021102531 A AU2021102531 A AU 2021102531A AU 2021102531 A4 AU2021102531 A4 AU 2021102531A4
- Authority
- AU
- Australia
- Prior art keywords
- wind
- aerial
- drone
- deep learning
- drones
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
- 238000000034 method Methods 0.000 title abstract description 48
- 238000013135 deep learning Methods 0.000 title abstract description 30
- 238000004364 calculation method Methods 0.000 title abstract description 13
- 238000004422 calculation algorithm Methods 0.000 abstract description 13
- 238000012549 training Methods 0.000 abstract description 12
- 238000013528 artificial neural network Methods 0.000 abstract description 9
- 230000007613 environmental effect Effects 0.000 abstract description 7
- 238000012544 monitoring process Methods 0.000 abstract description 7
- 238000011161 development Methods 0.000 abstract description 5
- 238000010801 machine learning Methods 0.000 abstract description 5
- 238000011160 research Methods 0.000 abstract description 3
- 238000007689 inspection Methods 0.000 abstract description 2
- 238000010606 normalization Methods 0.000 abstract description 2
- 238000013459 approach Methods 0.000 description 16
- 238000012360 testing method Methods 0.000 description 11
- 238000005259 measurement Methods 0.000 description 8
- 238000012417 linear regression Methods 0.000 description 6
- 230000003068 static effect Effects 0.000 description 6
- 238000007637 random forest analysis Methods 0.000 description 5
- 230000001419 dependent effect Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000007123 defense Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 238000011497 Univariate linear regression Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 231100001261 hazardous Toxicity 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000135 prohibitive effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
- G08G5/003—Flight plan management
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/08—Adaptations of balloons, missiles, or aircraft for meteorological purposes; Radiosondes
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
- G08G5/0073—Surveillance aids
- G08G5/0091—Surveillance aids for monitoring atmospheric conditions
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B64—AIRCRAFT; AVIATION; COSMONAUTICS
- B64C—AEROPLANES; HELICOPTERS
- B64C39/00—Aircraft not otherwise provided for
- B64C39/02—Aircraft not otherwise provided for characterised by special use
- B64C39/024—Aircraft not otherwise provided for characterised by special use of the remote controlled vehicle type, i.e. RPV
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W2001/003—Clear air turbulence detection or forecasting, e.g. for aircrafts
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Atmospheric Sciences (AREA)
- Environmental & Geological Engineering (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Automation & Control Theory (AREA)
- Indicating Or Recording The Presence, Absence, Or Direction Of Movement (AREA)
Abstract
THREE-DIMENSIONAL WIND, AIRSPEED CALCULATION, AND
PREDICTION METHOD FOR AERIAL DRONES USING DEEP
LEARNING
Abstract:
Aerial drones are well known as Unmanned Aerial Vehicles (UAVs) are effectively used in many
contemporaneous applications namely traffic monitoring, surveillance and reconnaissance, parcel
delivery, wireless internet facility-based drone delivery, and environmental and desertification
monitoring, etc. The aerial drone is operated either by humans or autonomously so that the aerial
drone can navigate to the respective location to collect the informational data in many applications.
The ubiquity and the versatile nature of the UAV (unmanned aerial vehicle) which is commonly
known as the drones are made to measure and calculate the wind conditions. The calculation of
the local wind's direction and speed becomes a vital process as it involves the drone's aviation.
This gets more important in UAV as it becomes important in its applications like inspection of the
infrastructure, monitoring the environment, and operations that involve search process and rescue
process. The drones involved in civil applications are commonly small and tend to get diverted
when the force of the wind is strong. These advancements either involve the experimentations
using the indoor environmental conditions that are controlled or involves complex computation
using mathematical representations. Still, there is more difficulty faced in the process of
computation of the wind speed. Thus, research and experimentation are made using the machine
learning algorithm. This invention is intended for the development of the method that measures
the speed of the air or wind around aerial drones. This method uses the deep learning neural
network for making the computation. The sensors are used for sensing the information regarding
the wind speed that is stored and processed using the deep learning algorithms.
1
UNKNOWN WIND SPEED
AERIAL DRONE
DATA
FEATURE VECTOR FEATURE VECTOR
NORMALISATION
ESTIMATED
WIND SPEED
FOR
AERIALDRONE
TRAINING
DATABASE
Figure 3: Calculation method for estimating the speed of the wind.
2
Description
Figure 3: Calculation method for estimating the speed of the wind.
Description
Field of the Invention:
This invention is intended for the development of the method that measures the speed of the air or wind around aerial drones. This method uses the deep learning neural network for making the computation. The sensors are used for sensing the information regarding the wind speed that is stored and processed using the deep learning algorithms.
Background of the Invention:
The ubiquity and the versatile nature of the UAV (unmanned aerial vehicle) which is commonly known as the drones are made to measure and calculate the wind conditions. The calculation of the local wind's direction and speed becomes a vital process as it involves the drone's aviation. This gets more important in UAV as it becomes important in its applications like inspection of the infrastructure, monitoring the environment, and operations that involve search process and rescue process. The drones involved in civil applications are commonly small in size and tend to get diverted when the force of the wind is strong. Thus, the computation of the wind speed and direction becomes a dominant task. The smaller the drones, the lesser the inertia thus these UAVs are prone to wind attack. The unmanned aerial vehicle is of two types namely 1. Multi-rotor UAV, and 2. Fixed-wing UAV. In the recent two decades, the multi-rotor UAVs have been used in most applications like research, defense, and commercial fields where their unique advantages are used when compared to the fixed-wing UAVs. The multi-rotor UAV performs better in the process of vertical take-off and landing. These advancements either involve the experimentations using the indoor environmental conditions that are controlled or involves complex computation using mathematical representations. Still, there is more difficulty faced in the process of computation of the wind speed. Thus, researches and experimentation are made using the machine learning algorithm.
Pramod et al discussed the forbidden challenges aroused during the flight performance of multi rotor small unmanned aerial vehicles (UAVs). They even made a comparative analysis on the various techniques in measuring wind speed and airspeed for multi-motor UAVs. Flow sensors, anemometers, and tilt-angle-based techniques are the three different sensor-based techniques discussed in this review. The simulation output shows the effective wind disturbances for the above-mentioned techniques. During the simulation process, power spectral density (PSD) functions, computational fluid dynamics (CFD), and probabilistic models are utilized. The proposed open-source python-based wind turbulence model is interfaced with an ultrasonic anemometer through effective embedded coding.
Patrick et al evaluated the real-time wind on a micro unmanned aerial vehicle using its inertial measurement unit. Wind speed and wind direction are determined with the obtained measurement data with the support of the onboard sensors. They only utilized the sensor where it is independent of any additional airspeed sensor and dedicated anemometer. They even evaluated the performance of the enhanced aerial vehicle via the wind tunnel and field tests process. The anemometer is placed in an open field nearer to the flight of micro UAV so that the anemometer can effectively gather the informational data about UAV to obtain a high degree of accuracy.
Liyang et al determined the rotary-wing wind speed for an unmanned aerial vehicle during the vertical take-off and landing. The usage of the airspeed sensor might severely affect the rotor's down-wash effect is the most challenging task to be performed. K Nearest Neighborhood learning based method is independent of the aerodynamic information. The offline training stage and online wind estimation stage are the two most essential stages included in the proposed system model. The airspeed of the aerial drone depends only on the data in an online estimation stage. The wind disturbance is obtained from the fan of the aerial drone during the quadrotor testing. In the hovering condition, the proposed algorithmic approach acquires a high degree of accuracy and robustness.
David et al estimated a wind velocity using two supervised machine learning approaches. Multi rotor copter attitude measurement in association with the machine learning approaches is implemented in the lowermost atmospheric region effectively tackles the wind velocity challenges. Onboard inertial drone data is accompanied by the wind tower measurement are obtained from the combination of the K-nearest neighbor (KNN) algorithm and a long short-term memory (LSTM) neural network. The efficient stabilization performance of the two hovering drones in a windfield for future prediction is possible. In comparison with these two approaches, the LSTM is the most appropriate method used in predicting wind conditions. LSTM is performed under the altering wind regimes with the support of the demonstrated linear regression model. The proposed complex and specific dynamic models are involved in the multi-rotor-based wind speed estimation process in a prohibitive environment like mountain terrain or off-shore sites.
Amer et al developed autonomous drone navigation using a Deep Convolutional Neural Network. Aerial drone autonomous navigation uses the onboard camera to visually obtain the predefined input about the navigation path. The aerial drone follows the steering commands provided by the combination of the deep Convolutional Neural Network (CNN). This enhancement of the drone is adaptable in the contemporaneous deployment scenarios using the augmented data for the multiple auxiliary navigation paths. Environmental and desertification monitoring, parcel delivery, and drone-based wireless internet delivery are the various application fields that use automatic drone navigation. The proposed algorithmic approach is completely based on the GPS map that relates the spoofing and enables navigation instead of human intervention for the operation purpose. They tested the drone model in two different scenarios.
Adrian et al made a literature review on unmanned aerial vehicles based on deep learning applications. Deep learning has an effective performance on a wide range of robotic applications in the areas of perception, planning, localization, and control. Autonomous robotic applications are extremely suitable for contemporaneous environmental conditions such as security, disaster rescue, and surveillance, etc. The limitation and performance of the UAVs using the deep learning approaches are reviewed in this paper.
Magdalena et al developed an internal stabilization system to measure the wind speed for quadrotor drones. They proposed a theoretical structure with the desired equations for the quadrotor control and the parameter required for the wind speed estimation is obtained from the static thrust test. The laminar wind condition is site tested along with the quadrotor hovering which is placed next to the 2D ultrasonic anemometer. The wind speed ranges between 0-5 m/s where the root mean square error (RMSE) value ranges between 0.26-0.29 m/s for the corresponding wind speed in a respective wind direction. The proposed system approach is low in cost.
Bian et al determined the three-dimensional wind based on ultrasonic sensor array and multiple signal classification. The wind speed and their direction is the most challenging task to be estimated to effectively control the wind turbine process and permits the linkage of wind power into the electrical grid. They accurately measured the wind speed in the three-dimensional (3D) space using the proposed system model. The wind speed and the corresponding direction of the wind in the 3D space are measured simultaneously using the semi-conical ultrasonic sensor array. Multiple Signal Classification algorithms are utilized to determine the noise and wind speed that influences the ultrasonic signal being transmitted in between the sensors. They determined the accuracy range in terms of root mean square error and mean absolute error. This systematic approach is considered to be the standardized method in determining the wind speed under the varying signal-to-noise ratio.
Objects of the Invention:
1. This invention is intended for the development of the method that measures the speed of the air or wind around aerial drones. The aerial drone collects the meteorological and atmospheric data to have an effective performance. Wind disturbance is the most sensitive and challenging factor to achieve the stable flight performance of the aerial drone. 2. This method uses the deep learning neural network for making the computation. The deep learning algorithm mainly responsible to read the data being represented which enables the extraction system inherent to the deep learning algorithm. 3. The sensors are used for sensing the information regarding the wind speed that is stored and processed using the deep learning algorithms. The information data from the environment is extracted from the sensor data.
Summary of the Invention:
Aerial drones are well known as Unmanned Aerial Vehicles (UAVs) are effectively used in many contemporaneous applications namely traffic monitoring, surveillance and reconnaissance, parcel delivery, wireless internet facility-based drone delivery, and environmental and desertification monitoring, etc. The aerial drone is operated either by humans or autonomously so that the aerial drone can navigate to the respective location to collect the informational data in many applications. The autonomous navigation of aerial drones along the predefined path is achieved using onboard sensors such as Inertial Measurement Units (IMU) and Global Positioning System (GPS). The navigation path is achieved through the closed-loop drone navigation system using IMU where the location is traced via the GPS.
The aerial drone is usually categorized into two different types namely, fixed-wing UAVs and multi-rotor UAVs. Among these two types, multi-rotor UAVs are the most commercial method used in the defense markets mainly due to their distinct features in comparison withfixed-wing UAVs. The vertical take-off and landing are possible via the multi-rotor UAVs. The usage of small unmanned aerial vehicles (sUAVs) is applicable for non-military applications due to their low cost. The aerial drone collects the meteorological and atmospheric data to have an effective performance. Wind disturbance is the most sensitive and challenging factor to achieve the stable flight performance of the aerial drone. The wind speed of the unmanned aerial vehicle (UAVs) is estimated using the airborne wind-measuring system which enables the ideal performance for various applications. The wind condition in specified locations is identified using the anemometer. To mount the anemometer, the fixed-wing UAV is more applicable than a multi-rotor drone. The wind speed estimation is more important for the effective performance of the aerial drone motion concerning the time-varying parameter algorithmic approaches.
The ultrasonic sensor such as the vortex and the time-of-flight (TOF) is utilized to determine the wind speed in the three-dimensional space. The proportional relationship between the average flow velocity and the eddy current frequency. The wind speed is estimated according to the acquired speed components and positional relationship among the sensors. An arc ultrasonic sensor array determines the wind speed and direction in a three-dimensional space. The aerial drone can take high-quality aerial photographs, and video via which it collects a large amount of imaging data. The 3D map model is developed using high-resolution images. The aerial drone is included in the rescue team during hazardous situations. The architectural model of the unmanned aerial drone consists of various layers namely, social layer, reflective layer, deliberative layer, executive layer, reactive layer, and physical layers. The deep learning-based algorithmic approaches include the above-mentioned layers in classifying the systematic applications with unmanned aerial vehicles.
The operation of the aerial drone is operated with relatively minimal experience due to the support of advanced control technologies. The aerial drones are relatively low in cost and hence, these are accessible to a wide range of operations. The area is accessed via the navigation of the drone in a desirable direction as per the estimation of airspeed. The sensor collects the raw data to solve the complex problem via the implementation of deep learning approaches. The deep learning approaches consist of various layers to predict the object in the navigation path of the aerial drone. The deep models including the computer vision tasks such as speech recognition, natural language processing, and signal processing are efficiently performed. The main reason for the drastic increase in the utilization of aerial vehicles due to the automation capability, low cost, and versatility. The deep learning implemented in the enhancement of the aerial drone consists of the components about the perception, guidance, navigation, and control of unmanned rotorcraft systems.
The wind speed calculation is possible in three-dimensional space is determined using various advanced technologies such as Artificial Intelligence, Machine Learning, and Deep Learning approaches to identify the current situation. The sensor collects the factor such as temperature, and air pressure which affects the wind speed and wind direction. Due to changing environmental conditions, the warm and cold wind varies the temperature of the air. The rate of changes in the pressure rate is essential in determining the wind speed and the airspeed rapidly increased due to the centripetal force. The wind direction is altered due to the rotational of the earth on the axis.
Detailed Description of the Invention:
The aerial drone is enhanced with 3D robotics is generally used in many contemporaneous applications. The aerial drone is provided by the propeller with the respective size and pitch and these propellers are equipped with four brushless motors. Accelerometer and Gyroscope, Accelerometer and Magnetometer, Barometer, and Gyroscope are enclosed within the flight controller. The aerial drone is maintained in a respective height based on the wind speed estimation along with the Global Positioning System (GPS) and barometric altitude measurement. On-board sensors collect the informational data and store these data in a logfile which can be accessed and analyzed during post-flight.
Aerial drone's three-dimensional wind and airspeed calculation are determined from the momentum theory. In aviation, the body of the aerial drone is fixed at the air velocity of va
[va, va, va]. The obtained air velocity is applied to the wind triangle va = 9 - vw from where wind velocity v is estimated. The velocity and speed of wind from the ground levelvg is acquired from the GPS drone data. Drone mass, total rotor disk area, local temperature air density, propeller inertia moment, and coefficient of lumped linear drag, are essentially important parameters to be initially identified for the effective performance of the aerial drone. The rotational acceleration and velocity of the propeller, motor torque, and accelerometer data are utilized to obtain the unknown parameters namely three-dimensional wind velocity and airspeed, vertical induced velocity, horizontal drag force, and vertical thrust from the following equations,
-=-H (1) m m
H -T c(vx + va&) (2)
T= 8pAtotvzU (3)
U IV - Va (4)
Pa = PT + PH (5)
PT = T(v' - v') (6)
PH - v (7)
PM = 1 Pa + Pr (8)
Here, Figure of Merit (FoM) is well known as efficiency factor that relates the motor mechanical power (Pm) with the aerodynamic power (Pa) and this FoM is obtained from a static thrust test.
Pm TW (9)
Pr =IfOO (10)
The atmospheric pressure is obtained using air density p at a temperature of 5°C. The rod rotating around its center is simplified to obtain the propeller moment of inertia I. The sensor utilized in the aerial drone effectively measures the accelerometer data directly. During the static thrust test, the velocity of the rotational propeller, propeller acceleration, and the motor torque are determined.
Static thrust test utilizes thrust stand of RC benchmark series of 1580, and dynamometer. Motor rotational velocity, battery voltage, current drawn from the battery, pulse width modulated remote control input, motor torque, thrust are measured during the static thrust test. The aerial drone is flight-tested with the reference static anemometer measurements to predict the coefficient of lumped linear drag. The three-dimensional anemometer acquires the high resolution at 1 wind direction with the wind speed rate of 0.01m/s. In the flight testing process, care must be taken. The various sensors are used in estimating informational data at the various point is obtained to calculate the wind estimate.
The proposed system includes the parameter for the effective measurement of three-dimensional wind and airspeed. Data collection, pre-processing, followed by splitting of data to undergo the training and testing process, and finally, the three-dimensional wind and airspeed of the aerial drone are identified are the steps included in the proposed system model. The meteorological data is obtained as input raw data which is being utilized in the pre-processing stage via the sensors. To attain the standardized model for the effective performance of aerial drones the unnecessary missing data are removed. From the deep learning approach, three different models are utilized and they are Linear Regression, Random Forest, and Deep Neural Network. The acquired data in the further process has been split up into training and testing data. Finally, the three-dimensional wind and airspeed are obtained against the actual values through the error values.
The linear regression model for the aerial drone is essential in attaining the relationship among the independent and dependent variables. The acquired data is directly linked to the linear regression model where it presumes the liable distribution of dependent variables. The value of the wind speed is nothing but a dependent variable. Wind direction, temperature, and pressure are considered as the independent variable. Since, the speed of the wind is associated with this independent variable and hence, the dependent variable is highly influenced by the independent variables in the case of multivariate regression instead of univariate regression.
The wind speed of the aerial drone is estimated using the below-mentioned expression,
y= _PO +pi11+ p2x2 +---pnxn+ (11) Here PO is the representation of the parameter; y is the wind speed, xi is the independent variables, and the error is denoted ass. The difference between the actual and predicted value is known as error.
The aerial drone includes the training model via the utilization of the random forest technique. The collection of the numerous data may increase the complexity and hence, the random forest efficiently identifies the data on each step of the process autonomously. This process continues still the final value is obtained. Deep Neural Network is also an effective method of determining the desired output. It consists of the number of layers where the information is passed on from one layer to another. The training process for the hidden units is continued until the final output is obtained. The weight is assigned for each layer where the desired output is nothing but the sum of the weights. The information data from the environment is extracted from the sensor data. The deep learning algorithm mainly responsible to read the data being represented which enables the extraction system inherent to the deep learning algorithm.
The deep learning approaches with the linear regression model achieve better performance gains. The proposed model includes the training and validation ratio with high precision information. The deep learning model performance is improved with the training and validation data to attain a sustainable amount of reasonable error. The most significant factor includes the ideal prediction of the temporal synchronization among the drone attitude data and wind speeds. The anemometer produces the attitude that comes in relationship with the wind velocity based on the wind gust, wind direction, and eddy-front information is transmitted to the anemometer. The training process in the proposed model uses the high-resolution data to provide an effective alignment procedure in between the drone data and recorded wind speeds. The aerial drone flight is performed with the persistent flows in a respective direction.
Claims:
This invention involves the development of the method for the calculation of the wind speed and its direction has the following claims:
1. This invention is intended for the development of the method that measures the speed of the air or wind around aerial drones. The aerial drone collects the meteorological and atmospheric data to have an effective performance. Wind disturbance is the most sensitive and challenging factor to achieve the stable flight performance of the aerial drone. i. From claim 1, the wind speed is estimated according to the acquired speed components and positional relationship among the sensors. ii. From claim 1, the aerial drone can take high-quality aerial photographs, and video via which it collects a large amount of imaging data. The 3D map model is developed using high-resolution images.
2. This method uses the deep learning neural network for making the computation. The deep learning algorithm mainly responsible to read the data being represented which enables the extraction system inherent to the deep learning algorithm. iii. From claim 2, from the deep learning approach, three different models are utilized and they are Linear Regression, Random Forest, and Deep Neural Network. The acquired data in the further process has been split up into training and testing data. iv. From claim 2, the three-dimensional wind and airspeed are obtained against the actual values through the error values.
3. The sensors are used for sensing the information regarding the wind speed that is stored and processed using the deep learning algorithms. The information data from the environment is extracted from the sensor data. v. From claim 3, the aerial drone includes the training model via the utilization of the random forest technique. Deep Neural Network is also an effective method of determining the desired output. The weight is assigned for each layer where the desired output is nothing but the sum of the weights.
THREE-DIMENSIONAL WIND, AIRSPEED CALCULATION, AND 13 May 2021
Diagram 2021102531
Figure 1: Architecture of the aerial drones and their corresponding layered structure.
Figure 2: Block diagram of the proposed system.
Figure 3: Calculation method for estimating the speed of the wind.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2021102531A AU2021102531A4 (en) | 2021-05-13 | 2021-05-13 | Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2021102531A AU2021102531A4 (en) | 2021-05-13 | 2021-05-13 | Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
AU2021102531A4 true AU2021102531A4 (en) | 2021-07-01 |
Family
ID=76584632
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
AU2021102531A Ceased AU2021102531A4 (en) | 2021-05-13 | 2021-05-13 | Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning |
Country Status (1)
Country | Link |
---|---|
AU (1) | AU2021102531A4 (en) |
-
2021
- 2021-05-13 AU AU2021102531A patent/AU2021102531A4/en not_active Ceased
Similar Documents
Publication | Publication Date | Title |
---|---|---|
González-Rocha et al. | Sensing wind from quadrotor motion | |
Johansen et al. | On estimation of wind velocity, angle-of-attack and sideslip angle of small UAVs using standard sensors | |
Tian et al. | Wind sensing and estimation using small fixed-wing unmanned aerial vehicles: A survey | |
Xiang et al. | Wind field estimation through autonomous quadcopter avionics | |
Corbetta et al. | Real-time uav trajectory prediction for safety monitoring in low-altitude airspace | |
Iwaneczko et al. | A prototype of unmanned aerial vehicle for image acquisition | |
Larrabee et al. | Wind field estimation in UAV formation flight | |
Donnell et al. | Wind characterization using onboard IMU of sUAS | |
Aleksandrov et al. | Optimization of lift force of mini quadrotor helicopter by changing of gap size between rotors | |
CN113821044A (en) | Bridge detection unmanned aerial vehicle autonomous navigation and stability control method based on reinforcement learning | |
US10876920B1 (en) | Auxiliary aerial vehicles for flow characterization | |
Vijayanandh et al. | Conceptual design and computational investigations of fixed wing unmanned aerial vehicle for medium-range applications | |
AU2021102531A4 (en) | Three-dimensional wind, airspeed calculation, and prediction method for aerial drones using deep learning | |
Tian et al. | UAV flight test evaluation of fusion algorithms for estimation of angle of attack and sideslip angle | |
Ariante et al. | Velocity and attitude estimation of a small unmanned aircraft with micro Pitot tube and Inertial Measurement Unit (IMU) | |
Lee | Helicopter autonomous ship landing system | |
Chmaj et al. | The dynamics influence of the attached manipulator on unmanned aerial vehicle | |
US11948467B2 (en) | System, device and method for measuring atmospheric conditions | |
Tian et al. | UAS-based wind estimation using sinusoidal gust model | |
Khaghani et al. | Evaluation of wind effects on UAV autonomous navigation based on vehicle dynamic model | |
Lyu et al. | A novel integrated navigation system based on the quadrotor dynamic model | |
Muhammad et al. | Total least squares estimation of aerodynamic parameter of micro coaxial helicopter from flight data | |
Qu et al. | Real-time wind vector estimation for a micro UAV | |
Rhudy | A dynamic model-aided sensor fusion approach to aircraft attitude estimation | |
París i Bordas | Control and estimation strategies for autonomous MAV landing on a moving platform in turbulent wind conditions |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FGI | Letters patent sealed or granted (innovation patent) | ||
MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |