RFQ / BOM 0 Sign In / Register

Select Your Location

Key technologies of UAV autonomous control system 4

July 16, 2021

2207

Compared with the flight control system of manned aircraft, the autonomous control system of advanced UAV is undoubtedly much more complicated. Judging from the current technical level and development status, the "automation" of the UAV control system has solved the problem of automatic flight control, but it is far from solving the problem of intelligent autonomous control. For its key technologies, there are many domestic and foreign related studies and reviews. You can focus on the classic document "The Role of Autonomy in DoD Systems" released by the US Department of Defense in 2012, which summarizes the six core technologies related to the autonomy of unmanned systems, namely perception, Planning, learning, human-computer interaction, natural language understanding and multi-agent collaboration, and based on the cognitive level view gives the technical status and challenges, so far the analysis and refinement of the key technology of autonomous control still has a strong guiding significance.

image.png

Based on existing research and practice, it is not difficult to find that most of the unresolved core issues of autonomous control are concentrated at the decision-making level, the organization level and the coordination level. Therefore, facing actual flight and mission scenarios, combined with capability requirements, the key technologies of the autonomous control system can be summarized as follows. It should be noted that the key technologies described below are not independent of each other. They are closely related, and even overlap and merge locally. They should be considered as a whole in research, application and practice.


4.1 Design technology of autonomous control system architecture


The main task of the UAV autonomous control system architecture is to connect the various subsystems into a whole, manage and dispatch each subsystem in a unified manner, so that each subsystem can complete the overall task in unison. The quality of its design is directly related to the UAV. The overall performance of the system and the level of intelligence.


The target and application scenarios of autonomous control technology are more complex and dynamic than previous automatic control systems. From the perspective of efficiency, the future working mode of UAVs will cover single-aircraft actions and multi-aircraft coordination modes. Many elements should be considered comprehensively when designing the system architecture, including the decomposition of the mission of the entire fleet into the specific goals of each drone, online mission planning, online optimization of the mission route of the formation, trajectory planning and tracking, and differences in the formation. Coordination between UAVs, realizing reconstruction control and fault management while taking into account the uncertainties of the environment and their own failures and damages. Therefore, it is necessary to design a reasonable system architecture to properly solve the problem of the division of software and hardware functions and the coordination of various elements in the system to ensure that the complex system is flexible, open, and configurable.


In response to such requirements, the currently widely accepted solution is to adopt a hierarchical design mode, which divides the system into multiple functional levels such as tasks, decision-making, and execution, so as to ensure the design of high-level functions such as tasks and decision-making and the bottom layer. Control execution and other restrictions are decoupled. On the basis of ensuring the fixed input and output relationship between layers, the functional design of each layer can adopt a flexible form. In the design of the UAV autonomous control system architecture, centralized or distributed communication/decision logic can be selected, and the universal bus can be used to connect various subsystems.


4.2 (multi-source) information collection, processing and fusion technology


Correct, reliable, timely and effective information is the prerequisite for the planning, decision-making, management and control of the autonomous control system. The collection, processing and fusion technology of uncertain information from multiple sources not only directly supports the realization of the drone's perception ability, but also the basis for the realization of various other capabilities. It plays an indispensable role in the use environment of complex information, highly confrontational, and changeable tasks. The important role of lack.


Under normal circumstances, the process of multi-source information collection, processing and fusion can be simply described as first obtaining relevant information through sensors or information exchange channels from multiple sources, and then combining and processing the obtained information according to certain criteria to achieve the The structured representation of data information, so as to obtain the available information related to the platform itself, environment, goals, situation, threats, etc. Therefore, for the realization of the autonomous control system of UAVs, the front-end multi-source information collection, processing and fusion technology is strictly a technical field with a wide range, mainly involving the following related sub-technologies: advanced sensor realization technology (such as time/ Displacement, navigation/positioning, detection/detection and other sensors); sensor signal processing technology (correction, compensation, noise reduction, time-space synchronization, etc.); autonomous navigation, positioning and timing technology (including relative navigation); unstructured unknown environment perception and Modeling technology; target detection, recognition, and tracking technology; enemy/we behavior understanding and intention recognition technology; information interaction and sharing technology; situation assessment/threat estimation technology, etc.


The challenges faced by multi-source information collection, processing and fusion technology mainly come from the uncertainty of information, that is, the information obtained under complex and confrontational conditions may be unstructured, incomplete, noisy, asynchronous, and unpredictable. , Even deceptive. In addition, the data form of the acquired information often presents high-dimensional, massive, dynamic and other characteristics, which also puts forward high requirements for the real-time nature of the acquisition, processing and fusion technology under the condition of drones.


4.3 Online real-time planning and autonomous decision-making technology


One of the important features of autonomous control is that it can solve complex planning and decision-making problems under uncertain conditions. The online real-time planning and autonomous decision-making technology of the UAV autonomous control system is mainly oriented to the realization of flight management and task management related functions.


Planning is essentially an optimized compromise between existing capabilities and target costs. The planning problem in the UAV autonomous control system mainly refers to task planning, which generally includes task priority sorting, multi-task assignment, trajectory planning, and task load. Planning, communication topology planning, and plan planning for system guarantee and emergency. Under normal circumstances, planning can be divided into offline global planning and online real-time re-planning. At present, most researches are on trajectory planning technology and task allocation technology, both of which have achieved a series of good results and are gradually moving towards engineering applications.


Decision-making is actually a selection process among feasible schemes, which requires effective reasoning, evaluation, and prediction based on the acquired information and existing knowledge, so as to obtain the final result. Decision-making in UAV autonomous control system mainly refers to autonomous behavior decision-making, such as tactical maneuver decision-making in autonomous air combat.


On the one hand, the challenge of online real-time planning and autonomous decision-making technology comes from the complexity and uncertainty of the environment and tasks, on the other hand, it comes from the complexity brought about by the expansion of the system application scale. For example, in distributed decision-making for UAV collaboration or cluster control, each agent not only needs to consider the environment and the agent’s own model, but also needs to consider the strategies that other agents may adopt when making decisions. The problem has a highly complex strategy space, and as the number of individuals and the scale of the planning problem increase, its complexity grows exponentially, and it quickly becomes difficult to solve.


4.4 High-precision/robust/adaptive/fault-tolerant control and execution technology


High-precision/robust/adaptive/fault-tolerant control and execution technology belongs to the category of traditional flight control field, and is mainly applied to the execution layer of hierarchical structure. Among them, high-precision instruction tracking control technology is a necessary guarantee for UAVs to achieve flight and missions, especially in some specific mission scenarios, such as landing/landing, air refueling (refueling), formation flying, and obstacle avoidance flying. Etc., the control accuracy will play a decisive role in the execution of the task; the robust control technology is mainly aimed at the uncertainties in the controlled objects (such as structural uncertainties and parameter uncertainties, etc.) to ensure the robustness of the system Stability and robust performance; adaptive control technology can face the wide-range dynamic changes of the control object and provide satisfactory flight quality. In traditional flight control, linear control methods and gain tuning strategies are used to solve the problem. The main research in the future And development directions include nonlinear adaptive control technology for large envelope flight, task adaptive control technology for task changes, and large variable configuration adaptive control technology for variant aircraft; fault-tolerant control technology is the system's fault-tolerant ability The basis of realization, fault tolerance is mainly reflected in the self-diagnosis ability of faults and the ability to reconstruct faults during the online operation of the system. Therefore, the main research content of fault-tolerant control technology includes fault detection/diagnosis/isolation, self-repairing reconstruction control, etc.


In addition, control actuators with high reliability, high bandwidth, and high power-to-weight ratio are also the key to the realization of the UAV autonomous control system. In-depth research is needed on the corresponding servo actuation technology, such as electro-hydraulic servo valve technology and direct drive valve technology. , Hydraulic sealing technology, electro-hydrostatic servo actuation technology, electromechanical servo actuation technology, comprehensive evaluation and verification technology of actuator durability, etc.


4.5 Task-oriented autonomous collaborative control technology


Collaboration is a task-oriented high-level intelligent activity. In the case of complex/hard tasks and limited abilities of executing individuals, complex tasks can be effectively completed through collaboration and cooperation between individuals. Task-oriented autonomous collaborative control is the embodiment of the advanced autonomous capabilities of UAVs. The autonomous collaborative control technology of drones mainly solves the control of multiple drones and the collaborative behavior between humans and drones, and realizes the collaborative tasks of manned-unmanned platforms and the collaborative tasks between multiple unmanned platforms. In addition to various uncertainties, the challenges it faces also need to solve the difficult problems of handling distributed collaborative decision-making, management and control. At this stage, the research on task-oriented autonomous cooperative control mainly focuses on two directions: manned-unmanned cooperative control technology and formation cooperative control technology.


In the direction of manned-unmanned cooperative control technology, the main problem is how to match the capabilities of the man-machine system, which requires autonomous control with variable permissions, so that the control permissions can be dynamically transferred between the pilot, auxiliary systems, and autonomous systems. . At the same time, it is necessary to solve the problem of human-machine cooperative behavior control, which mainly includes: mathematical modeling and reasoning judgment of human intervention/intention, the basic structure design of cooperative behavior controller and its multi-loop control stability analysis.


In the research direction of formation cooperative control technology, the main research contents include formation design and maintenance, formation dynamic adjustment and transformation, formation flight control, formation reconstruction control, formation collision avoidance/obstacle avoidance, etc. For example, in formation flight control, common implementation methods include: Leader-Follower method, behavior-based method, virtual structure method, consistency method, etc.