01 traffic monitoring

A promising application of UAV in security mission is to strengthen traffic supervision system. At present, traffic supervision system has been widely deployed and has become an important part of intelligent transportation system (ITS) infrastructure.



        Although the system is very important, it is less deployed in many rural areas, exists only in specific locations, can only carry out simple traffic counting, and can not be used for comprehensive traffic operation, mainly considering the impact of cost and benefit.

In this regard, UAV provides an economic and effective means to meet the needs of rural traffic supervision system. Highway traffic flow has certain dynamics and uncertainty, so it is necessary to provide real-time and accurate information in accessible and remote areas.

1. Congestion monitoring

In recent years, traffic congestion has become increasingly serious. People can improve the control and response of traffic facilities by collecting real-time information of traffic conditions, so as to effectively reduce travel delays, and improve medical and health conditions by shortening the time for the wounded to receive rescue services. At a specific time, the great value of the monitoring network comes from only a small part of the monitoring network.

Unfortunately, the high value part of the monitoring network is constantly changing and usually unknown. For example, the location of vehicle congestion due to traffic accidents depends on the location when unpredictable events occur. The traditional traffic supervision system ensures the ability of the whole traffic network to respond quickly under changing conditions by deploying fixed position detectors (including cameras) with high density. When information outside the range of these fixed detectors is required, manual evaluation is required.

The national traffic flow remote sensing Alliance (ncrst-f) has recognized that using UAVs to provide aerial view and rapid response to transportation operations is a low-cost method.

Useful information collected in UAV traffic monitoring includes vehicle lane change frequency, average vehicle distance, number of heavy vehicles, accident type, vehicle trajectory and type. Although loop detectors can be used, they only provide local information and cannot provide details such as vehicle lane change.

On the other hand, the UAV equipped with camera can provide the global view and relevant information of expressway, and enhance the real-time monitoring ability of road.

In addition, UAVs have more advantages than manned aircraft. For example, UAV can fly at a lower altitude; UAVs can also be used when weather conditions are not suitable for manned aircraft flight. This kind of application should solve two problems when collecting information through UAV: keep the road in the field of view of the camera; Process images and collect relevant data.

Usually, the information contained in the traffic data captured by UAV is much more complex than that obtained by traditional monitoring system. UAV video includes not only traditional data such as average speed, density and flow of traffic flow, but also horizontal data of each vehicle, such as vehicle trajectory data, lane change data and vehicle tracking data on the road.

In addition, the video frame captured by UAV contains multiple vehicles, and the video frame rate is very high, so the amount of data to be processed will be very large. Considering these characteristics, data collection, restoration and analysis are regarded as an important part of big data analysis in transmission. The problems that must be solved include physical layer problems, communication problems and network layer problems.

The project is studied as a case of UAV in remote sensing and multimodal transport. The main objectives of the project are as follows.

Develop reliable software and hardware architecture, including coordination and response components for UAV autonomous control.

Develop sensory platform and sensory interpretation technology, focusing on active vision system to realize real-time constraints of sensory data processing.

Develop effective reasoning and algorithm technology to obtain dynamic and static geographic, spatial and temporal information related to the operating environment.

Develop plan, predict and record identification technologies to guide UAVs and predict and deal with the behavior of ground vehicles.

Development of simulation, specification, project, verification technology and modeling tools based on complex environment and function.

2. Driving behavior monitoring

In the study of driving behavior, detailed and accurate vehicle trajectory data are also needed. Driving behavior model captures the driver’s maneuver decision under different traffic conditions, which is an important part of micro traffic simulation system. Compared with the traditional traffic supervision system, the image captured by UAV has a certain challenge to detect and track vehicles.

First, the cameras on the UAV monitoring platform change frequently. Because the camera on the UAV may rotate, move and flip during video recording.

In addition, sudden vibration may also occur due to wind fluctuations, which may cause negative effects in vehicle tracking.

On the other hand, accurate trajectory data of each vehicle is required in driver behavior research models (such as vehicle following and lane change models). Lack of vehicle data and tracking errors may affect the accuracy of model parameters.

Therefore, high-resolution images are very important for accurate calculation of vehicle speed and lateral position in the process of vehicle detection and tracking. Vehicle recognition methods can be divided into optical flow and feature extraction and matching methods.

Abnormal driving behavior has been used to identify drunk driving (DWI) and prevent related accidents. At present, law enforcement officers rely on visual observation to detect such behavior and identify potential drunk drivers. However, this method is limited by human error and is limited to a small range of vehicles.

In order to overcome these limitations, UAVs can be used to monitor driving behavior, prevent accidents and improve highway safety, and effectively and timely detect and analyze dangerous driving activities on highways. Law enforcement officials often use these bad behaviors to identify potential drunk drivers summarized by the National Highway Traffic Safety Administration (NHTSA) in 2010.

In order to observe eight potential adverse behaviors, six key indicators must be identified and quantified. These key indicators include vehicle ID, speed, forward distance, lane change frequency, lane change time and acceleration. Therefore, calculating these key indicators is the primary task to determine computer vision algorithms and quantify bad behavior. Calculating the six key indicators requires the position of the vehicle relative to the lane line and the recognition and tracking of the vehicle through video frames.

02 nuclear, biological and chemical accidents

Nuclear, biological and chemical (NBC) accidents are another example of cooperation between manned and unmanned systems. In post disaster recovery projects, it is usually necessary to map and quantify the pollution of radioactive elements in areas that are difficult or impossible for people to reach. This task is an ideal application scenario for robot system. In particular, UAV provides a promising general solution.

The UAV program can be used to investigate the radioactivity at the accident site during emergency response, and also to inspect the structure of the nuclear reactor shell. However, the UAV platform is difficult to use at the nuclear site. Because UAVs usually rely on GPS to achieve stability and control, GPS is unreliable or ineffective near or inside metal clad buildings. The data captured by UAV is also difficult to measure the pollution quantitatively.

In another emergency reconnaissance case, the UAV should fly into toxic clouds and bring back pollutant samples for analysis.

Because human operators do not want to be close to the accident site, UAVs need to have the ability to fly for a long time and resist strong upwind. It needs to estimate the concentration of smoke plume produced by gas source in the atmosphere. The detection of source location has a variety of applications related to environment and search and rescue missions. Real time plume concentration estimation makes it possible to locate the gas source location and deploy countermeasures against the adverse effects of plume.

With the help of gas transport model, fixed, surface or air sensors in the plume area can be used to solve the problem of plume estimation. This procedure uses the initial assumptions at the specified location to measure the concentration and source location to build a possible source location map.

Sensor movement control supports the same detection by collecting the measured values of sensors, such as environmental measurement. When the sensor is installed on the UAV platform, the most important thing is to plan the sensor path to achieve low working time or low energy consumption of the UAV, avoid obstacles in the unstructured dynamic environment, or collect target information efficiently.

03 search and rescue

Search and rescue (SAR) has many forms, including urban search and rescue, field search and rescue, maritime search and so on. Each form has corresponding risks and poses risks to victims and search and rescue personnel.

1. Urban search and rescue

Urban search and rescue (USAR) is defined as the strategy, strategy and operation of location search, and provides medical and rescue services for victims.

USAR is an area where UAVs may play a role. Before rescuers enter, they can determine a way to deal with search and rescue. USAR has at least 7 difficulties in the rapid rescue of collapsed buildings.

With the development of urban search and rescue mission, in order to obtain the best methods and lessons, it is necessary to objectively evaluate and analyze the effective experience of the mission.

Rescue technology includes casualty assessment, monitoring and development and optimization of rescue tools.

Regional participation establishes a collaborative relationship between professional rescue coordination and region based first responders to achieve the balance of professional knowledge and resources.

The information system can identify, collect and manage multiple data streams and convert them into information in order to provide better advanced planning and timeline situational awareness.

Technology integration involves the verification and integration of more advanced technologies.

Crisis management is a practical and scalable management system.

Available budgets affect purchases and the rapid deployment and use of systems and technologies. In the long run, it will have a significant impact on the development of new systems and technologies.

The USAR team can perform the following tasks:

Physical search and rescue in collapsed buildings;

Provide emergency medical assessment and care for the trapped;

Assess and control hazards, such as gas or power services;

Assess and reinforce damaged buildings.

The task of the human UAV work team is to explore the disaster area and provide sufficient information for situation assessment. The human UAV rescue team consists of at least one UAV, several personnel located in the remote control room, and one or more human UAV operators. The team is geographically dispersed. The UAV team includes:

UAV operator, operate the UAV on site

UAV mission experts watch UAV video streams and guide UAV operators to perform tasks

Measures to protect the safety of UAV team

During deployment, UAV mission experts mainly work with UAV operators to provide additional perspectives for UAV operators. The UAV team then evaluated the video data. The information obtained from the video data is directly provided to the national fire brigade and is also used for subsequent unmanned ground vehicles (UGV) to perform tasks.

For situational awareness, this requires that this method can integrate different views on the environment and get different views and needs. In order to achieve this goal, UAV needs more autonomy to sense the environment and cruise autonomously. However, the disaster area is a place with poor environment. This poses a challenge to the durability of UAV systems operating in urban environments.

In addition, the stability of radio link broadband cannot be guaranteed in disaster areas. When mobile systems operate in harsh environments, the limited availability of computing resources and low-quality sensors also pose a great challenge to the autonomy of UAVs.

These mission characteristics require that UAV can be modular and flexible in sensor and planning capability. UAVs must be able to operate in unstructured indoor and outdoor environments, such as collapsed buildings. The navigation system must be able to work without external auxiliary equipment (such as GPS), because its availability cannot be guaranteed.

Because there are local wind changes in this environment, UAV must also provide robust flight capability. A key feature of full autonomy in urban disaster areas is on-board processing and decision-making. Search assignment also requires UAV to have specific task identification function. Identifying and locating people, animals, or objects (e.g., landmarks, signs, or landing areas) is a core issue of the USAR mission.

2. Field search and rescue

Field search and rescue (wisar) often needs to search large areas in remote areas with rugged roads. Because the ground search robot needs to face the situation of large search area and limited mobility. Therefore, using small UAV to provide aerial images of search area for field search and rescue is an excellent alternative method.

If we want to successfully achieve extensive deployment, the UAV supporting field search and rescue needs to have the characteristics of portability, durability and easy operation. These requirements add many limitations, including those generated by selected specific UAVs; Restrictions caused by human factors, especially from minimum training requirements; Restrictions imposed by the control device used; Constraints from the specific task at hand, including the need to enter the existing team structure.

3. Maritime search

Using appropriate UAV system can search at sea more effectively. Marine SAR used as system integration (SOS) is a regional scenario to implement and demonstrate the architecture method. It uses a variety of systems, including UAVs, coordinated command and control systems, communication systems and other larger manned ships.

At present, various sensors and data sources are being used: coastal radar, patrol or surveillance aircraft, ship radar, civil aircraft or ship reports, etc. Each sensor has its own characteristics. For example, a coastal radar station continuously covers within its radar horizon, but fails completely outside the horizon.

For beyond line of sight and continuous naval surveillance, the most effective, economical and flexible way to observe the area of interest is to use surveillance UAV for regular surveillance. This method provides the ability to take advantage of the interdependence between all systems. Therefore, it is necessary to generate a robust, efficient, network centric architecture that can be generated.

Generating architecture for SOS is a multi-objective optimization problem with many variables and constraints. The information required to generate the schema is as follows.

The primary purpose of SOS is to select the coast guard search and rescue capability as the issue.

Stakeholders: the coast guard has a number of systems with different capabilities that can be used at multiple sites in the region. In addition, fishing vessels, civilian and commercial vessels can temporarily join the SOS to provide assistance in the event of a disaster. The coordination and command and control center guides the combination of manned ship and unmanned aircraft in operation.

Rough order of magnitude (ROM) budget.

Key performance attributes (kPa) of SOS: performance, enforceability, robustness, modularity and network centrality.

Maritime SAR with multiple UAVs faces several problems.

First, commercial UAVs have limited fuel capacity and therefore cannot operate indefinitely.

Secondly, the survival probability of survivors in a given area varies with time and wind, which is often called containment probability (POC).

Third, many UAVs and fuel station systems should be automatically controlled.

If the basic drift and related uncertainty are not properly estimated, it is still difficult to predict the drift and expansion of the search area. The direct method is to use an ammeter to measure the motion of an object relative to ambient water. The search and rescue optimal planning system (sarops) uses an environmental data server to obtain wind and current forecasts from multiple sources. The system gives the search paths of multiple search units to greatly improve the detection probability of search increment.

When reconstructing particle dispersion based on observed or modeled vector field, ocean diffusivity is an important factor. In many cases, a simple stochastic model is sufficient to estimate the dispersion of SAR objects in a relatively short time period. Regional estimates of dispersion and integration time scales (possibly seasonal) should be carefully considered because they may have a great impact on SAR object diffusion.

Tagged:

Leave a Reply

Your email address will not be published. Required fields are marked *