Talk about drones and their key Technologies (Civil)

 

Drones first appeared in the 1920s, and in 1914, in the First World War, someone developed a small aircraft that was not piloted by humans, but controlled by radio. Modern warfare is the fundamental driving force behind the development of UAVs. The world's first unmanned aerial vehicle was born in 1917, and the real use of unmanned aerial vehicles began in the Vietnam War, mainly for battlefield reconnaissance.

Unmanned aircraft referred to as "unmanned aerial vehicle", English abbreviation for "UAV", is mainly the use of radio remote control equipment and self-provided program control device to control the unmanned aircraft.

The Origin of drones

 

Drones first appeared in the 1920s, and in 1914, during the First World War, someone developed a small aircraft that was not piloted by humans, but controlled by radio. Modern warfare is the fundamental driving force behind the development of UAVs. The world's first unmanned aerial vehicle was born in 1917, and the real use of unmanned aerial vehicles began in the Vietnam War, mainly for battlefield reconnaissance.

In 1982, Israel and Syria in the Bekaa Valley war, Israel used UAVs for reconnaissance, jamming, and trapping, and the role of UAVs was once again valued and developed.

In the Gulf War of early 1991, drones became a "must have" battlefield capability, with six pioneering unmanned aerial systems engaged. Provides high-quality, near-real-time, all-day reconnaissance, surveillance, target acquisition, interdiction, and battle damage assessment.
 

The Kosovo war was the largest and most effective use of drones in any war.

In the first Russian Chechen anti-terrorism war in 1995 and the second Russian Chechen anti-terrorism war in 1999, the Russian army used unmanned reconnaissance aircraft to conduct reconnaissance and surveillance in the combat zone, especially in the second Chechen War, the Russian army's "bee" unmanned reconnaissance aircraft to detect a large number of rebel data, providing accurate data for the Russian army's precise strikes.

In the aforementioned wars, UAVs mainly played the role of reconnaissance. In the Afghan war, the United States used Predator as a carrier and launched AGM-114C and Helfa air-to-ground missiles, which realized the direct fixed-point attack of UAVs launched missiles in actual combat for the first time and further developed the function of combat UAVs. It is also the verification of the actual use of unmanned combat aircraft, and the real start of unmanned warfare.

Types of drones
 

Functional UAVs can be divided into military UAVs and civilian UAVs.
 

 

Civilian UAVs include inspection and inspection UAVs, communication relay UAVs, remote sensing mapping UAVs and time-sensitive target strike UAVs. This paper focuses on civilian unmanned aerial systems.

 

Control station: including display system, control system;

 

Communication link: including airborne communication and ground communication.

1. Flight control system

 

The flight control system is the "pilot" of the UAV, which is the core system of the UAV to complete the whole flight process such as take-off, flight, task execution and return recovery.

Flight control generally includes three parts: sensor, on-board computer and servo actuation equipment, and the functions achieved mainly include UAV attitude stabilization and control, UAV mission equipment management and emergency control. Among them, a large number of sensors assembled in the fuselage (including angular rate, attitude, position, acceleration, altitude and airspeed, etc.) are the basis of the flight control system and the key to ensure the accuracy of aircraft control. In the future, UAV sensors are required to have higher detection accuracy and higher resolution, so a large number of new technologies such as hyperspectral imaging, synthetic aperture radar, and ultra-high frequency penetration are applied in high-end UAV sensors.

 

The existing flight control system is a combination of open source and closed source systems. Domestic excellent UAV manufacturers, in order to improve the specialization of the system, most of them evolved their own closed-source systems on the basis of open source systems. Compared with open source systems, UAV manufacturers' own closed source systems have added many optimization algorithms, simplified the tuning harness, and become more simple and easy to use.

 

2. Navigation system

 

The navigation system is the "eye" of the UAV, and the combination of multiple technologies is the future direction.

The navigation system provides the position, speed and flight attitude of the reference coordinate system to the UAV, and guides the UAV to fly according to the specified route, which is equivalent to a navigator in the man-machine system.

 

At present, the navigation technology used by UAVS mainly includes inertial navigation, positioning satellite navigation, terrain aided navigation, geomagnetic navigation, Doppler navigation and so on.

 

The unmanned aerial vehicle navigation system is mainly divided into two types: non-autonomous (GPS, etc.) and autonomous (inertial guidance), but they are vulnerable to interference and error accumulation, respectively. The future development of unmanned aerial vehicles requires obstacle avoidance, material or weapon delivery, automatic approach and landing and other functions, which require high precision, high reliability and high anti-interference performance. Therefore, the combination of multiple navigation technologies "inertial + multi-sensor +GPS+ photoelectric navigation system" will be the direction of future development.

3. Power system

 

At present, civil industrial UAVs are mainly oil-powered, and consumer UAVs are mainly electric.

 

Different uses of UAVs have different requirements for power devices. Low-speed and medium-low altitude small UAVs tend to use piston engines, low-speed short-range and vertical take-off and landing UAVs tend to use turboshaft engines, and small civilian UAVs mainly use electric motors, internal combustion engines or jet engines.

 

The turbine is expected to gradually replace the piston, and the new energy engine will increase the endurance.

 

Professional-grade UAVs are currently widely used as piston engines, but the piston type is only suitable for low-speed and low-altitude small UAVs. With the continuous improvement of turbo-engine thrust-to-weight ratio, life expectancy and lower fuel consumption, turbines will replace pistons as the main power model of UAVs.

 

New energy motors such as solar energy and hydrogen energy are also expected to provide more lasting power for small drones.

4. Data link system (communication system)

 

The data link system (communication system) is the bridge between the drone and the control station, and is where the real value of the drone lies.

 

The uplink communication link is mainly responsible for sending and receiving remote control commands from the ground station to the UAV.

 

The downlink is mainly responsible for sending and receiving telemetry data, infrared or television images from the drone to the ground station.

 

Most common UAVs use customized line-of-sight data link, while mid-altitude and long-endurance UAVs use over-the-line-of-sight satellite communication data link.

 

The development of modern data link technology promoters UAV data link is developing in the direction of high speed, broadband, security and anti-interference. With the on-board sensors, the precision of positioning and the complexity of the task continues to rise, the bandwidth of the data link put forward strong requirements, the future with the rapid development of high-speed processors on board, it is expected that the transmission rate of the existing RF data link will double in a few years, and the future may also appear laser communication methods.

Key foundational technologies for smart drones

Drone "vision" technology

 

One of the key technologies in making drones "smart" is the ability to perceive their surroundings through machine vision and translate the results into data that can be passed on to other applications via an OS (operating system).

 

At present, the mainstream machine vision hardware technologies in the field of UAV are: binocular machine vision, infrared laser vision, ultrasonic assisted detection and other ways.

 

1. Binocular machine vision

 

Binocular machine vision is based on the principle of triangulation, which is similar to the principle of human eye's restoration of the three-dimensional world. By comparing the Angle difference of the same object in the picture taken by two cameras in the same direction, the distance is determined, so as to restore the stereo model of the three-dimensional world from the two-dimensional image.

 

 

Binocular machine vision requires only two cameras, but it requires high computing power.

 

The threshold for binocular machine vision is not to calculate distance L from perspective information α, β and spacing d, but to enable the computer to "extract" objects from the background in the picture. At present, Qualcomm's UAV reference design that supports binocular machine vision uses the flagship chip Snapdragon 801/820, which shows its high requirements for computing power.

 

It is a natural thing for the human eye to distinguish an object from the background screen, but it is different for the computer: the image of the same scene in the camera image plane of different viewpoints will be distorted and deformed to different degrees, to let the computer blur the object, the image segmentation algorithm needs to do convolution/differentiation and other large computation; The overall computing performance of drones, which require real-time ranging scenarios, is even higher.

2, infrared laser vision

 

In order to avoid a large number of calculations in computer vision to identify objects and improve accuracy, a group of manufacturers represented by Intel use infrared laser vision technology, such as Intel RealSense machine vision module. Its basic principle is shown in the following figure, and its ranging principle is similar to binocular vision, but the recognition object is replaced by an infrared laser point hit on the surface of the object. This essentially eliminates the computational need for object recognition.

 

 

The necessary cost of infrared laser vision is to replace two cameras with infrared cameras and increase the hardware cost of an infrared laser scanner as well as power consumption. The infrared laser scanner consists of an infrared laser transmitter and a MEMS scanning mirror, and the overall hardware cost is higher.

 

In addition to the small computational requirements, infrared laser machine vision also has two major advantages: compared with binocular, its application time and range are wider, and it can be used in dark and poor lighting conditions indoors; Compared with binocular, it has higher ranging accuracy and can accurately restore the three-dimensional data of objects.

 

3, ultrasonic detection

 

Ultrasonic obstacle detection is a relatively mature technology, which has been widely used in various military/civilian applications.

 

The advantage of ultrasound is that it can effectively identify objects such as glass, wires and other binocular vision/infrared laser vision cannot be prepared for ranging.

 

The disadvantage is that the accuracy is poor, can only be used to detect the existence of obstacles, can not extract accurate spatial information for path planning.

 

Point hover technique

 

The core application of consumer UAV is based on the aerial photography function of UAV, and the aerial photography function requires the highest technical index of the UAV system is the flight stability.

There are several technical means used in hover positioning technology:

 

1) GPS/IMU combined positioning

 

2) Ultrasonic assisted height determination

 

3) Image-based optical flow positioning technology

 

GPS/IMU positioning technology

 

The principle of GPS/IMU positioning is a more traditional and mature positioning method.

 

GPS can measure the current horizontal position and altitude of the UAV, and the flight control system compensates and controls the UAV according to the deviation of the UAV position and altitude relative to the hovering point, so as to realize the fixed-point hovering.

 

However, the GPS signal update is slow, and the GPS signal is easy to receive interference, affecting the actual control effect. Therefore, in engineering practice, IMU information of aircraft is introduced to filter with GPS signal to obtain more accurate and updated position and altitude information. This mode can also ensure that only IMU can provide emergency position and altitude information when GPS is out of order, but the solution results are easy to diverge when only IMU information is used to solve position and altitude. Therefore, this method is only suitable for hovering control in the open outdoors, and is not suitable for indoors or in the environment with signal shelter.

 

Ultrasonic assisted high technology

 

The ultrasonic ranging sensor is a mature ranging sensor, which can measure the distance of obstacles in front of the ultrasonic sensor according to the time difference between the ultrasonic signal and the return. When the UAV is equipped with a downlook ultrasonic sensor, a more accurate distance from the ground can be measured, thus assisting in the realization of fixed height control. However, ultrasonic assisted height fixing has no effect on drift control of horizontal position.

 

 

 

Optical flow localization

 

Optical flow positioning is a technology that uses image sensor to analyze the image captured by the sensor and indirectly calculate its own position and motion information.

 

With the evolution of image processing algorithm and the development of image processing hardware platform, the accuracy and real-time performance of this algorithm are guaranteed, so that it can be applied in unmanned aerial system.

Optical flow positioning is a method to find the corresponding relationship between the previous frame and the current frame by using the changes of pixels in the time domain and the correlation between adjacent frames, so as to calculate the motion information of objects between adjacent frames. In general, optical flow is caused by the movement of the foreground object itself in the scene, the movement of the camera, or a combination of the two.

In the application of UAV, the body of UAV is equipped with an optical flow camera to the ground, and positioning is carried out according to the observed ground image, the principle of which can be understood by the following figure: When the UAV moves relative to the ground, the picture captured by the earth observation lens will "move" in the opposite direction. According to the height of the UAV from the ground (which is also the reason why the optical flow sensor and the earth ultrasonic sensor are paired) and the amount of pixel movement in the earth observation image, the distance of the UAV relative to the ground can be calculated.

When the UAV uses optical flow positioning technology to determine its own position, it can use a general control algorithm to achieve horizontal and altitude positioning. The current optical flow technology can basically achieve a stable hover in the indoor environment, but over time, there will still be a drift of more than ten centimeters to tens of centimeters. However, this low-frequency, small-amplitude position change is acceptable for aerial photography.

 

Tracking technique

For aerial drones, a new trend is to use the tracking shooting mode, that is, to set a point of interest to the drone, the drone will automatically track and shoot the point of interest, which is the development trend of UAV intelligence.

The current tracking shooting technology is mainly divided into two types:

1) GPS tracking;

2) Image tracking.

GPS tracking

GPS tracking technology is relatively simple, that is, the person being tracked needs to hold a remote control, and obtain the satellite positioning information of their current position, and then send this information to the drone, the drone to receive the target position as the target, and navigation.

GPS tracking is a relatively rudimentary method of tracking, and most drones on the market use this method.

Image tracking (including face recognition tracking)

Image tracking technology means that the UAV can track the target completely according to the image characteristics of the set interest points, which involves image recognition and image tracking of the target object, especially in the target movement scene, where the image background changes relatively and the target shape changes greatly, the accurate tracking of the target needs to apply deep learning technology. It is a hot research direction of artificial intelligence at present.

Automatic obstacle avoidance technology

The flight safety of UAVs has always been the core issue related to the large-scale commercial application of UAVs. How to sense obstacles and avoid obstacles autonomously is the most advanced research topic in the field of UAVs flight safety.

With the large-scale commercial application of autonomous flight and tracking flight of UAVs, the requirement of autonomous obstacle avoidance in the process of autonomous aerial photography and tracking has become more urgent.

At present, there are three different obstacle avoidance technologies:

1) Obstacle avoidance based on ultrasonic detection;

2) Barrier technology based on LiDAR;

3) Realsense monocular + structured light detection and obstacle avoidance.

Ultrasonic ranging obstacle avoidance

This technology is similar to the traditional reversing radar system, according to ultrasonic detection, to learn the distance information of obstacles, and then adopt the corresponding strategy to avoid obstacles, which is characterized by a close detection distance and a small detection range, but the method is very mature and easy to achieve.

Binocular vision avoidance

This technology is based on binocular vision image depth of field reconstruction method, depth of field reconstruction of the scene in the field of view, through the depth of field information to judge the obstacles in the field of view, the detection range is wider, the distance is longer, the corresponding security is higher, but the technical difficulty is large, and will be affected by the change of light intensity.

Obstacle avoidance technology based on LiDAR

This technology relies on the laser radar technology widely used in unmanned cars to scan the environment around the drone and conduct map modeling.

Realsense Monocular + Structured light detection obstacle avoidance

RealSense is an earlier vision perception system released by Intel. It uses the "active stereoscopic imaging principle", which mimics the "parallax principle" of the human eye, by punching out a beam of infrared light, tracking the position of the beam with the left infrared sensor and the right infrared sensor, and then using the triangulation principle to calculate the "depth" information in the 3D image. With a depth sensor and a full 1080p color lens, gestures, facial features, foreground and background can be accurately recognized, allowing the device to understand human movements and emotions. According to Intel, Realsense has an effective range of up to 10 meters.

Wireless image transmission technology

One of the core technologies of UAV aerial photography is wireless image transmission, and the transmission capability is an important factor to measure the UAV aerial photography capability.

Drone aerial photography technology

Drone aerial photography technology can simply be divided into 2 points according to the literal "drone" + "aerial photography" :

1, image shooting technology, that is, imaging and image processing technology; For example, the number of pixels, aperture size, etc., but there are many parameters that affect the image quality on the camera module: single pixel size, sensor technology, lens group technology, ISP technology, etc.

2. Uav platform technology mainly refers to the fuselage control technology that provides a stable aerial photography environment for aerial photography.

Image shooting technology: At present, the image shooting program on the market is the integrated application of several major brands of mainstream camera modules, UAV manufacturers do not have much technical space in this respect, and because the technology development has been relatively mature, the gap between different product programs is not large.

Uav on-board platform stabilization technology: In addition to the UAV's own flight technology such as flight navigation and control, the relevant technology for the UAV to achieve stable aerial photography platform guarantee. This technique is the most critical factor affecting the quality of the image.

The significance of a stable shooting platform:

When shooting video, the picture shake and tilt will seriously affect the fluency and beauty of the picture;

When taking photos, especially in low light, if the exposure time is long, the shaking of the fuselage will cause the blur of the picture; If the exposure time is reduced, the sensitivity needs to be increased, and the noise increases, affecting the picture quality. Therefore, the stability of the fuselage is crucial for filming.

 

Main factors affecting fuselage stability:

According to the typical situation of the current four-rotor UAV, several factors that cause disturbance to the fuselage position and attitude can be summarized as follows:

1. The horizontal position and height caused by inaccurate hover positioning;

2, the body tilt and shake when the body is operating;

3. Interference caused by motor vibration and sudden wind. Different strategies are adopted to deal with different types of disturbance.

For horizontal and height drift, in the outdoor, that is, when the GPS signal is good, the unmanned vehicle mainly locates according to the GPS signal. However, due to the limited accuracy of the civilian GPS system itself and the low update frequency, it is difficult to simply rely on the GPS system for positioning, and usually the UAV will introduce the inertial module for combined positioning.

When indoors or in the case of limited GPS signal reception, the UAV system also uses the ground camera for optical flow positioning. Optical flow positioning is a new image-based positioning method which is developed in recent years. It can be used well when it is close to the ground.

If the drift of the position is a slow dynamic disturbance, then the body tilt and jitter caused by the UAV maneuver are high-frequency disturbance factors, which have a significant impact on the picture.

When the UAV needs to move the position, the attitude of the four-rotor fuselage must be greatly adjusted, especially when the maneuver has just occurred, the fuselage attitude has been adjusted by 40 degrees.

Disturbance factors such as body tilt when the fuselage moves in the horizontal direction and jitter when the body moves have a great impact on the image shooting effect, which must be offset by mounting a stable head.

For the motor vibration, sudden wind disturbance and other factors, considering that it is a relatively high frequency disturbance, the hollow rubber ball spring can be used to filter out the high frequency vibration, which can achieve better results. Due to the large randomness in form and size, it is difficult to completely eliminate the influence, so it can only be considered to restrain the influence by combining various forms such as the head and optical flow.

 

Finally, one technology that cannot be ignored is electronic image stabilization technology. Electronic image stabilization technology is to feel the movement of the body through the sensor without the aid of mechanical equipment, so as to crop and splice the image on the display screen. From the perspective of software, the intention of image stabilization is realized to a certain extent.

Pinion technology

The head head plays an important role in suppressing the disturbance of the fuselage, such as active tilt and passive interference, which affect the aerial photography effect.

Generally speaking, the airborne head is usually a three-axis head. As shown in the following figure, the "three axes" of the three-axis head is divided into three axes, pitch, yaw and roll, also known as three degrees of freedom, and they are respectively controlled by a motor. In other words, the camera can realize the decoupling of the three degrees of freedom with the UAV through the motor control on the frame of the three-degree-of-freedom head (the value of the UAV's three degrees of freedom: pitch, yaw, and roll), and play a role in isolating and offsetting the influence of the UAV movement.

Three-axis PTZ technology mainly includes parts: 1, motion sensitivity; 2. Offset control.

Motion sensitivity: The camera part installed in the innermost layer needs to be able to sense the attitude deviation of the camera. A three degree of freedom gyroscope is usually installed.

Offset control: that is, when the camera is sensitive to deviate from the set attitude (generally a horizontal state), the reverse motion is applied through the motor to offset the motion change.

From the above point of view, the accuracy and frequency of the sensor, as well as the accuracy and power of the motor output, and the performance of the control algorithm have a relatively large impact on the final effect. However, from the current product technology point of view, as long as the UAV equipped with a three-axis head in aerial photography, there is basically no big difference in the use of experience.

In terms of function alone, the key factors are: 1. The isolation degree between the head and the fuselage; 2. Controllable Angle range of PTZ; 3, response speed; 4, the level of accuracy.

The importance of the head for aerial photography

Displacement compensation: Even if the better GPS+ optical flow positioning technology is adopted, the UAV will still drift relatively large, the amplitude is about 0.3m, when the displacement occurs, the picture center will have a mechanical head in order to further ensure the stability of the picture, it is necessary to introduce a mechanical head to stabilize the picture.

Through simple geometric calculation, it can be seen that when the distance between the photographed object is relatively close, the horizontal drift has a greater impact on the picture. However, when the distance is far from the object being photographed, the effect is less. At this time, only the smaller degree of tilt of the head can correct the picture deviation, so that the photographed object can return to the center of the picture.

Attitude compensation: Compared with the position movement of the UAV, the disturbance of the UAV's own attitude has a more severe impact on the picture. When the distance between the photographed object is close or far, the influence is greater.

Ultra remote control drone technology

You can sit in front of your computer, and with the click of a mouse, you can send a drone to another airspace or even a country.

The principle is actually very simple, the whole system needs to have two 4G access points, one access point on the drone, the other on the controller. The PC sends instructions to the drone through the wireless network connection to control the flight path of the drone, while the drone sends high-definition video captured by the built-in camera to the user, and the user can adjust the flight path of the drone while monitoring the surrounding environment.