Unmanned Aerial Vehicle classification , Applications and challenges : A Review

In past few years, unmanned aerial vehicles (UAV) or drones has been a hot topic encompassing technology, security issues, rules and regulations globally due to its remarkable advancements and uses in remote sensing and photogrammetry applications. This review paper highlights the evolution and development of UAV, classification and comparison of UAVs along with Hardware and software design challenges with diverse capabilities in civil and military applications. Further, safety and security issues with drones, existing regulations and guidelines to fly the drone, limitations and possible solutions have also been discussed.


Introduction
Technological amelioration has impacted significantly social, economic and personal life, from business approaches to international wars.These transformations can be visualized by getting benefited from these technological advancements.Unmanned aerial vehicle (UAV) also known as remotely operated aircraft is the best example to visualize the change.Unmanned Aerial vehicles do not need any pilot onboard and can be operated autonomously or remote pilot control [1,2].UAV is an integral part of the Unmanned aerial system which incorporates UAV, communication link and ground control station.UAV overcomes the limitation of the terrestrial system in terms of accessibility, speed and reliability [3].UAV can provide cloud-free and high-resolution images to serve the commercial applications such as agriculture, mining and monitoring.UAV was originated in defense for reconnaissance and combat purpose.Perhaps in 1916 first ever semiautomatic aero plane was developed (aerial torpedo).In 1933 Royal Navy used the drones for the gunnery practices.Later on with the advent and integration of advanced navigation sensors UAV became an integral part of armed forces.The emergence of technology not only removed the limitations of UAV exercises in the military but expanded their wings in commercial applications related to agriculture, scientific activities, recreation, servile, delivering goods, photogrammetry and many more [4].Agriculture and Infrastructure cover the maximum share of unmanned aerial vehicle applications.The utilisation of autonomous UAV in agribusiness is reaching out at an energetic pace in Crop health monitoring, early cautioning frameworks, forestry, fisheries, and wildlife protection [5,6].Onboard Narrow band visible, near-infrared and thermal sensors administer the variability of the field from nutrient mapping to crop health monitoring, so that farmer makes smart farming decisions on time.

Comparison of Satellite, Aircraft and UAV
Satellite and Aircraft based technologies are the conventional methods of remote sensing which have their benefits and hindrances in the form of Coverage area, spectral, spatial and temporal resolution with onboard sensors.Along with less revisit frequency, satellite suffers from cloud cover conditions where information beneath the clouds is shadowed entirely.Aircraft systems may sustain with higher payloads and speed, but the hovering at a place and maintaining high and low speeds are the challenges [7].Several studies were performed by various researchers to compare the three technologies such as NDVI survey of larger farmland, soil and crop mapping, Land classification.Unmanned Aerial Vehicles somehow bridges the gap of conventional technologies regarding hovering capability, maintain low and high Speed, low elevation and more importantly higher spatial and temporal resolution of images.
A comparison of satellite, aircraft and UAV is depicted in table 1.

Based on Aerodynamics
A variety of UAV system has been developed and in the advancement phase, some of them includes the Fixed-wing aircraft [9][10][11], chopper [12,13], multi-copter [14], motor parachute and glider [15][16][17], UAV with Vertical takeoff and landing [18][19][20], congregating ready-made parts [21] and commercialized UAV [22,23].all of them are specified for a specific mission and have their zeros and ones.Fixed wing drones are very simple but saturated in designing and manufacturing, because of successful generalisation of larger fixed-wing planes with slight modifications and improvements.Fixed wings are the main lift generating elements in response to forward accelerating speed.The velocity and steeper angle of air flowing over the fixed wings controls the lift produced.Fixed wing drones require a higher initial speed and the thrust to load ratio of less than 1 to initiate a flight [24,25].If fixed wing and Multirotor are compared for a same amount of payload, fixed-wing drones are more comfortable with less power requirement and thrust loading of less than 1.Rudder, ailerons and elevators are used for yaw, roll and pitch angles to control the orientation of aircraft.Figure 2 shows the force applied on fixed-wing aircraft.

Figure 2: Fixed Wing UAV Aerodynamics
Fixed wing drones cannot hover at a place, and they cannot maintain their low speed.Subsequently, it can be seen that lift to drug ratio denotes the lift generated by a wing counter to drag generated.Fixed wing drones are more compatible with larger L/D ratio and with higher Reynolds number.Unfortunately, fixed-wing drones are less noticeable for L/D <10 for the reason that Reynolds number and efficiency decreases for smaller drones.
Flapping wing drones are primarily inspired by insects such as small hummingbirds to large dragonflies [26,27].The lightweight and flexible wings are inspired from the feathers of insects and birds which demonstrate the utility of weight and flexibility of wings in aerodynamics.However, these flapping wings are complex because of their complicated aerodynamics.Flapping drones can support stable flights in a windy condition, unlike fixed-wing drone.Light, flexible and flapper wings provide the flapper motion with an actuation mechanism.Intensive research on flapping wings has been carrying out by drone community and biologist because of their exclusive manoeuvrability benefits [28].Fixed/flapping-wing: Integrated effect of the fixed and flapping mechanism is used where fixed wings are used to generate lift whereas flagging wings are used for generation of propulsion [29].These type of drones are inspired by dragonfly which uses two pairs of wings in order to increase the lift as well as thrust forces.Hybridisation using fixed and flapping wing increases overall efficiency and aerodynamic balance [29].
Multirotor: Main rotor blade produces a forceful thrust, which is used for both lifting and propelling.
Multirotor unmanned aerial vehicles are capable of vertical takeoff and Landing (VTOL) and may hover at a place unlike fixed-wing aircraft [30,31].Multirotor are designed by number and location of motors and propellers on the frame.Their hovering capability, ability to maintain the speed makes them ideal for surveillance purpose and monitoring.The only concern with Multirotor is that they need more power consumption and makes them endurance limited.Abott equations are used for exact calculation of power and thrust requirements in multirotor aircraft.

Based on Landing
Horizontal takeoff and landing (HTOL) and vertical takeoff and Landing (VTOL): HTOL may be considered as the extension of fixed-wing aircraft.They have high cruise speed and a smooth landing.VTOL drones are expert in flying, landing and hovering vertically [33], but they are limited by cruise speed because of the slowing down of retreating propellers [32].

Based on Weight and Range
Some researchers and organisations have classified the drones based on weight and range.Table 2 presents the list of the unmanned aerial vehicle based on the weight and range.

Hardware Design and Challenges
The designing of the Unmanned aerial system includes the unmanned aerial vehicle and other subsystems which includes communication link between UAV and user, ground control station and accessories like gimbal, payload.The design of UAV itself integrates the parts evolving from vehicle frame to complete ready to fly the aerial vehicle.Selection of components like airframe, Controller, motor, propellers and the power supply is the crucial task and needs in-depth knowledge and fullfledged mathematical calculations to design a UAV for a specified mission.Figure 4 describes the subsystems and modules for the design of UAS.

Aircraft design
The design challenges of an aircraft rely on the type of application which specifies the coverage area, maximum altitude, speed, climb rate, flight time or endurance, and stability [39].All the specifications are prone to vary contingent on applications and the environmental effects.Higher altitude specifies a large coverage area and improves survivability although maximum altitude is limited by the aviation regulations.Climb rate also increases the survivability.Flight time is strictly dependent on the type of operation and aerodynamic design of aircraft.
The main components of aircraft subsystems are inertial measurement unit, motors, propellers and receiver, processor and an Airframe.The most common metallic materials to manufacture aircraft are alloys, aluminium and titanium, whereas nonmetallic materials include transparent and reinforced plastic [40].Multicopter have the N brushless motors with N propellers.Electronic speed controller serves their purpose by varying the power supplied to motor commanded from throttle stick.They can fly in a particular direction and adjust their elevation, i.e. pitch (along X-axis, heading of quadcopter), roll (along Y-axis) and yaw (along Z-axis) by taking the inputs from Inertial Measurement Unit (IMU) consisting of three-axis accelerometer, gyroscope which provides 3-axis raw data and a GPS unit.

Ground Control System
The typical ground station consists of a wireless router along with a computer to capture, process and display of data.Typically a ground control station should fulfil requirements such as open system architecture, compatible with different platforms like airborne, ship and ground, execution of data in real time, ability to control multiple UAVs, payload control and communication with other ground control stations [41,42].Other safety and a security function that can be expected from the ground control station include the warnings and emergency action plan in case of any failure, power outage restoration.

Data Link
It set up a communication channel between the Aircraft sensors and ground control station (GCS).A wireless link IEEE 802.11, is used to make a communication between aircraft central data unit and ground control station, for this purpose routers equipped with omnidirectional antennas with high gain can be used to minimise path loss and make a signal to noise ratio higher.Now a day's typical antennas work on 2.4GHz and minimum 12dBi gain.Additional wireless link based on orthogonal frequency division multiplexing (OFDM) is used for online video and images transmission to a ground station.

Accessories
With the advancement of drones, to carry out the applications such as photogrammetry, film shooting, mapping of a field, digital elevation models, monitoring and surveillance, UAV compatible cameras such as multispectral camera, thermal camera [43], hyper-spectral camera [44], digital camera [45] and film imaging units [46,47] were used.Frequently, cameras weighing less than 12 lbs.are preferred for FPV applications, and a minimum of 12 megapixel camera is required for agriculture applications.Table 4 describes the developments in multispectral sensors for UAV.

Data collection and analysis
The

Applications
Since Drones provide supremacy over conventional remote sensing technologies and their benefits lie in terms of less power consumption, less risk to human life, ease to data collection, hovering, and ultra-high spatial resolution forges them an excellent choice for surveying and mapping.Following pioneer, studies demonstrate the relevance and uniqueness of drones in the civil, logistics, agriculture and Defense sectors.Figure 5 depicts the potential applications of UAV in civil, environment and defence sectors.The sole aim of precision agriculture is to apply optimum amount of input at the right time and place to make better products.Common practices of precision farming are the data collection and variability mapping of agriculture lands, data analysis, making farming management decision based on results inferred from analysis and finally controlled application such as pesticide spraying and fertilisers.Agriculture has widely adopted the art of remote sensing using traditional satellite and aerial platforms.The capacity of a satellite to monitor map the vegetation is built upon the spatial, temporal and spectral resolution of sensors onboard such as MODIS, OLI, and AVHRR [48][49][50][51][52].In 2013, the launch of Operational Land Imager (OLI) on-board Landsat 8 with a temporal resolution of 30m provided images with a spatial resolution of 30meters and 15 meters for panchromatic in 11 bands ranging from 435 nm to 12510 nm.[53,54].A considerable amount of research work has been carried out in the area of precision farming using satellite-based remote sensing, and the results are auspicious.Research work including wheat yield variation due to climate variation [55], mapping of irrigated areas using AVHRR time series analysis [56], estimation of crop yield [57][58][59], Crop water stress management [60,61], forest cover classification and monitoring [62] has impacted Indian and global agriculture in terms of productivity analysis and farming management.However, this technique is somewhat limited to coarse resolution and cloud cover, Later on, UAS introduced the cheaper and low altitude alternative to providing high-resolution images.
[63] Used a Microdrone MD4-200 with a team ADC lite digital CMOS camera with image resolution of 1200x1024 pixels to estimate nitrogen and aboveground biomass of soybeans, alfalfa and corn crops.[64] captured the images of agriculture land using helicopter based unmanned aerial vehicle in a thermal band with 40nm resolution and 400-800nm spectral range with 20nm resolution using onboard Thermal and multispectral camera to target the biophysical parameters.[65] Used a drone along with multispectral and thermal sensors to delineate the spatial variability of water within a commercial rain-sustained vineyard.

Drones in Forestry, fisheries and wildlife protection
Drone bridges the gap of satellite imagery of less availability and cloud cover by performing Various tasks like measuring canopy height, tracking of forest wildlife's, support in forest management, forest fires detection and control, survey forests and mapping canopy gaps easily.[66] Targeted forest fire detection and monitoring using remotely controlled fixed wing UAV onboard thermal and hyperspectral sensors.Animal poaching and adverse climate conditions have a destructive impact on wildlife.UAV equipped with thermal sensors [67] along with satellite are being approached for monitoring, tagging and counting of animals which help to curb the poaching of animals and conserve the wildlife.

Drones in defence
The advent of UAV was started initially with the aim of transacting the war missions like intelligence, spying, reconnaissance vigilance and target detection; later they were introduced for civil and logistic applications [68].USA, UK, Russia, India and Israel are the leading countries in the development and deployment of military drones.In 2017 the acceleration in the proliferation of military along with civilian drones was observed, and a maximum number of drone strikes by USA and UK were noted.Breakthrough research and remarkable advancements in the area of swarming drones, jet-powered and Microdrones.According to the Bureau of Investigative Journalism, the US launched more than two times more strikes in 2017 than the year 2016.Following the bar chart shows the data of US strikes on Somalia, Yemen and Afghanistan from the year 2014-2017 [69].

Drones for civil applications
Drones are being fascinated in all commercial stratums from electricity companies to the railway industry.Electrical companies are preferring drones for inspection of high tension lines with ease of risky task of climbs and power outages [71,72].Railway companies have employed drones for monitoring and inspecting the track faults in constrained access areas.The Indian government is planning 3d mapping of thousands of kilometres long railway corridors and national highways.Amazon inspired from margarita pizza delivery with drones, tested the parcel delivery with drones although commercialisation of this project is awaited.Drones are helpful in performing search and locate operations of missing people during calamities condition.A trial to locate people in Donegal mountain range, Ireland and rescue operation of 200 people in flood zone by Chennai police, India exemplifies the potential and necessity of drones.Medical facility delivery using drones performed in many countries like the USA.Electricity generation through high elevation and high-speed drones is another exploratory area.There is always a tradeoff between Payload capacity and flight time in drone technology.Conventionally, onboard lightweight lithium-ion batteries are used to supply power to UAV, but their power backup is not comparable with other batteries.With the increase in payload, endurance decreases and hence mission may not get its completion.Fixed wing drones are efficient in power usage, but they have the drawback of hovering and speed control.Flying a single drone may also encounter a flight failure due to some technological and climatic reasons, so there is always a need to provide backup.Upcoming swarm flight of drones can execute this task, where in case of failure of one drone, others complete the mission.This technology is dependent on Swarm motion of insects, ants and birds and makes use of artificial intelligence, yet in developing phase.Drones are still limited by controlling through human operators, integration of Artificial Intelligence will allow a drone to make smart decisions and operate accordingly instead of human controllers.
Possible gains and harms are yet to be explored in this direction [73,74].Drones also suffer from windy conditions and adverse climate changes.Spraying Drones are efficient to spray less area, but for mass spraying, they become less efficient, and the operation cost becomes high.
Another concern is too technical learning of garners to make use of drone-based precision farming and to make the drone-based system fully automatic from image acquisition to making complex statistical models and decision support system.A GPS mounted on drone connects with four satellites to detect the position, velocity and elevation accurately.Since GPS signals are very much prone to noise and interference, there is a finite possibility of losing contact.At that time it is recommended that instead of emergency landing, their location should be estimated.The inertial navigation system combined with GPS provide a solution for this situation.Efficient algorithms have to be designed and tested to estimate the position and elevation correctly.
Besides the drone hardware design challenges, the cameras used for precision agriculture applications also puts some limitations.The multispectral images collection is very much prone to get affected by total irradiation along with sun angle and adverse weather conditions such as rain, heavy wind.Comparison of UAV data and satellite data puts two significant limitations that data has to be resampled to make equal spatial resolution images and secondly, if there is a cloud cover then it is almost impossible to compare the images since the information beneath the ground gets shadowed.
Onboard thermal sensors can detect the water utilisation of plants based on radiated temperature.The temperature variations in plants are exiguous which makes it difficult to discriminate other factors which may affect plant water such as sun irradiation.So further research is required.

Safety, Privacy and Security Issues
Some life-threatening issues because of rapid use of drones uncovered the concerns and challenges towards safety and security.Airworthiness, malicious practices and interference to public property are the primary concerns of safety, which puts a significant question mark on use of drone since modern approaches to resolve these issues are not up to the mark and does not give guarantee for the safe use of a drone.Air traffic control needs to be installed and monitored to ensure the inference avoidance in the airspace.

Controlling bodies and Regulations
Since long, UAV has been used for military purpose but the bans, inadequate rules and shortcomings interrupted the universal acceptance of UAV.The advancement and accelerated use of UAV in commercial applications demands for regulation challenges in order to ensure the safe, secure and authenticated use.[77].International Regulatory bodies like the International Civil Aviation Organization (ICAO) and European Aviation Safety Agency (EASA) invites states and organizations to frame policies and standards for civil aviation enterprise [78].Several countries are making efforts in this direction and have proposed regulations [79], some of them are being discussed in the following section.Australian Civil aviation safety authority revised the unmanned aircraft regulations in 2016 and included the new rules for remotely piloted aircraft.A pilot license and certificate is required for remote operation of UAV weighing greater than 2 kg.The certification covers many aspects including pilot information, maintenance, liability and safety aspects.Similarly, German air traffic act imposes rules and demands authorisation from aviation authority for an unmanned aerial vehicle weighing more than 5 kilograms and not being used for recreational purposes.Authorisation ensures the privacy, public safety and information protection.Unmanned aerial vehicles weighing more than 25 kilograms are restricted to take a fly beyond visual line of sight.France recently imposed two regulations for the use of civilian UAV.These regulations classified the UAV into three classes creational, experimental purpose and particular activities drone.Drone authorisation, altitude limits, weight and performance limits are mentioned in the regulations.France aviation laws also restrict the movement of the drone geographically such as in military peripheries, historic monuments, national parks and nature reserves.In India, Director these rules to all instructors, operators and manufacturers.Violation of licensing rules imposes the same penalties as applicable to manned aircraft.The United States Federal Aviation Administration (FAA) integrated the unmanned aerial system in the National Airspace system with the conditions of not compromising with safety, security and capacity.

Conclusion
The future generation is dependent on drones; they will create a new market.Drones are being upgraded and adopted by almost all commercial markets including precision agriculture, logistics and infrastructures.Future technology focused on increasing endurance, payload, improvement in the interaction between human and UAV and making clear rules and regulations for the safe and secure operation of UAV.Besides this, Integration of Artificial intelligence with drone technology will enable the drone to take decisions and independence to human controllers.

Figure 3 :
Figure 3: Classification of UAV based on Landing, Aerodynamics and weight

Figure 4 :
Figure 4: Unmanned Aerial System Subsystems growing popularity with endless possibilities of drones drive us to figure out the methods to data collection and analysis to utilize them with leveraged potential.Since collection and processing of UAV data is a long process, involving data collection, pre-processing of images, classification of images for features extraction, calculation of mathematical equations based on reflectances such as Indices and then the creation of a suitable model for results visualisation and interpretation.This section of the paper describes each step and makes a comparison of software's dedicated for UAV data collection and processing.Data collection step involves the flight planning, marking Ground control points (GCPs) for samples collected and the UAV flight itself.Correct Pre-processing with a needful accuracy of UAV data is the key step for the development of models because lack of pre-processing will distort the following processes and will generate the wrong outcomes.Pre-processing of UAV data includes the images selection, accurate georeferencing and ortho-rectification and mosaicking, (aligning images using Image control points, point cloud and GCPs).Classification of georeferenced UAV images includes the supervised and unsupervised techniques for planimetric features extraction like road, railway track, agriculture land, forest, land cover and water bodies.Maximum likelihood classification (MLC), random forest and Support vector machine (SVM) are the examples of relevant classifiers.Manual classification can be tried in case of unsatisfactory results.The next step is dependent on the case study, basically this type of processing is tried for Land use land cover classification, Agriculture biophysical parameters determination and soil study.Some of the popular Indices are Normalized difference vegetation Index (NDVI), Green-Red Vegetation Index (GRVI), Soil Adjusted Vegetation Index (SAVI) and Modified Chlorophyll Index (MCI).The last step is the creation of models for a specific case study.Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 27 November 2018 Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 27 November 2018 doi:10.20944/preprints201811.0601.v15.1 UAV Data Processing SoftwareUAV software simplifies the flight planning, processing and analysis of UAV data to provide actionable intelligence to service providers and farmers so that inputs can be optimised and better decisions can be made with reduced cost.Besides agriculture, these software are often utilised in mining, construction, surveillance, rescue operation and recreational purpose also.A good UAV software at least must include the automation of UAV flight plans, augmented view, Geo-rectification of images and 2D/3D models generation.Table6creates a comparison chart of various UAV software with their pros and cons.

Figure 7
Figure 7 Number of drone strikes carried out by the USA .preprints.org) | NOT PEER-REVIEWED | Posted: 27 November 2018 Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 27 November 2018 doi:10.20944/preprints201811.0601.v1 General of Civil Aviation (DGCA), Department of Civil Aviation drafts and regulates the policies for remotely piloted aircraft.A Unique Identification Number along with UAV operator Permit is required to fly drones, with proper adherence of guidelines such as prohibition to restricted areas such as Eco-Sensitive areas, beyond 500m into the sea from the coastline and beyond 25 km from international border [80, 81].Israel civil aviation authority regulates the laws regarding manufacturing, training and operations included the flight elevation, regulated and recognised routes and communication devices in Israel and enforced Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 27 November 2018 Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted: 27 November 2018 doi:10.20944/preprints201811.0601.v1

Table 1 :
UAV, Airborne and Satellite System Comparison

Table 4 :
Common Multispectral Cameras for UAVApart from multispectral cameras, thermal and Hyperspectral cameras are making the pace in remote sensing with drones.Drones equipped with thermal sensors are being used in mining, oil and gas industries.FLIR, telops commercial giants, have developed the thermal cameras compatible with drones.Hyperspectral sensors based remotes sensing records wavelengths with narrow spectral bands typically 5nm over the visible and NIR range.Hyperspectral images provide much more information with an ultra-high resolution where entire information is inherited in each pixel in contrast to multispectral sensors.Table5provides information about the developments in thermal and hyperspectral cameras for UAV.

Table 5 :
Representative Thermal and Hyperspectral Cameras

Table 6 :
UAV data processing software comparison

Table 7 : State of the art and Proposed solutions to UAV challenges
[75,76]