news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

author | li jun

edited by yang ruiqi and liu yukun

phoenix.com technology news: september 20, yesterday, the first low-altitude economic innovation and leadership conference "phoenix yufeng smart low-altitude" was held in fengtai, beijing. han wei, chief scientist of china shipbuilding group, delivered a keynote speech on the design and application of unmanned system architecture. in his speech, han wei said that although china's low-altitude economy is in its infancy, it should also start to consider the construction of system architecture.

han wei said that the development of foreign drone systems is very fast, with characteristics such as clustering, systematization, intelligence, low cost and modularization. algorithms, software and data are common in the field of new productivity, and their importance is increasing. they are not only important production factors for building the low-altitude economic system architecture, but also important outputs.

in general, there are many things that can be learned from the current development of low-altitude economy and military unmanned equipment. the future trend of both is to be based on national production, applicable to more unmanned platforms, support large-scale, distributed overall design, and develop in the direction of intelligence.

it is understood that the first low-altitude economy innovation leadership conference is themed "phoenix yufeng leading low-altitude smart development" and is co-organized by the beijing low-altitude economy industry development alliance and phoenix.com. the conference brings together top experts and scholars to discuss industry development trends, release the latest innovative achievements, and jointly draw a new picture of low-altitude applications through in-depth exchanges on cutting-edge science and technology.

the following is the full text of the speech, edited and published:

i am very honored to be entrusted by the conference organizer to share my work experience and insights. i am from china shipbuilding industry corporation. we have done some work in the field of unmanned equipment, especially in the system integration level, "mutual integration and learning of military technology and civilian technology". from this perspective, i would like to share my experience with you.

maritime unmanned systemsbasic development trends.

whether from the perspective of top-level planning requirements or system design, the application of unmanned equipment and unmanned systems at sea presents a "systematic application" state, which is similar to the "low-altitude economy" from this perspective. systematic application involves the design and planning of the entire system integration architecture, including policy, system, and technical planning. as a military research institute, we study how to better utilize the advantages and benefits of unmanned equipment, especially cluster equipment, at the technical level.

the report of the 20th cpc national congress pointed out that we should accelerate the development of unmanned intelligent combat forces, increase the proportion of new domain and new quality combat forces, and systematically promote the construction of new quality combat capabilities. by definition, the maritime intelligent unmanned system includes various unmanned platforms, unmanned boats, unmanned submersibles, underwater pre-positioned weapons and underwater drones at sea. it can be used independently on a single platform or integrated through autonomous networks and intelligent algorithms. it has a series of characteristics such as high emergence, self-organization, and cluster intelligence.

in terms of space, the maneuvering space is wider, and the adaptability and flexibility are higher. in terms of time, unmanned equipment has evolved from 3d to 4d, and has long-term, stable, and in-depth characteristics. at the same time, it has power advantages, including asymmetric advantages, continuous use advantages, and system advantages. the ukrainian war has profoundly reflected this, and there have been cases of manned troops surrendering to drones. these are the characteristics of unmanned equipment.

the basic components of the maritime intelligent unmanned system, in addition to the platform, also include the cluster overall control center and the external control system, which is similar to the so-called low-altitude control in terms of architecture. in addition, from a foreign perspective, foreign drone systems are developing very fast, with characteristics such as clustering, systematization, intelligence, low cost, and modularization. first, there is combat theory support, which is policy support in a sense. the central government encourages the development of low-altitude economy, which is called policy support. platform support means that we must have excellent aircraft and a large number of collaborative verification experiments and applications. the second is the huge data system, with too many things to design and relatively dense data; the third is intelligent drive, a large amount of data fusion and cross-industry fusion, the integration is relatively complex, involving multiple types of heterogeneity, and the timeliness requirements are relatively high.

from the perspective of architecture design, we pursue modularization, distributed computing, real-time optimization and intelligent decision-making, which are serious challenges. "architecture" is also being developed abroad. the first thing americans did was the control system architecture ucs for unmanned systems; another one was the joint architecture jaus for unmanned systems, which is for drones. the joint architecture of the u.s. military's drones officially released the universal unmanned system architecture for maritime unmanned systems - umaa last year. the characteristics of these architectures are that they go deep and integrate from the data model and design component framework to the modules inside the framework.

umaa defines four parts, platform interface, data model and management method. the platform interface defines the connection between platform and payload, platform and control station, manned and unmanned. here is a very interesting point. there are many drone platforms, and more can be used for offshore platform operations. how to optimize? optimize payload capacity and service capability? optimize control algorithm? these are all things that the architecture algorithm needs to consider. umaa defines eight high-level functions and their standard service interfaces, defines the interfaces of related software libraries, and defines service libraries, to maximize heterogeneous interconnection and selection of the best.

in general, first, maritime equipment is an important development direction in the future and needs to be accelerated; second, the high complexity poses severe challenges to the system architecture; third, it takes a long time to develop abroad, and both the military and civilians in china are doing it. for example, the navy has released unmanned equipment interoperability 2.0, and many things are still in progress. our low-altitude economy is in its infancy, and we should also start to consider it. when it reaches a certain stage of development and all the chimneys are up, it will be a bit late to strengthen the relationship at the system and system level by then.

design concept of maritime unmanned system.

what are the characteristics of unmanned equipment? we need to design unmanned architecture with unique features.

the first type of unmanned equipment includes drone equipment. the iteration of new technologies is very fast, which requires the architecture to have independent upgrade and efficient integration capabilities. this also includes current cars, which require a large number of ota, real-time online upgrades, data collation, etc. when designing the entire system architecture, we must consider integrability and deployability.

the second feature of low-cost and high-reliability equipment is modularity and high reuse, which are also the focus.

the third are algorithms, software and data, which are common things in the field of "new quality productivity" and are becoming increasingly important. they are not only important production factors but also important outputs.

the fourth is efficient heterogeneity and interoperability. we achieve this through three aspects: the first is full decoupling; the second is full abstraction, including hardware resource abstraction, service abstraction, and service-oriented architecture design; the third is autonomous interconnection and openness, which is our design method perspective.

the design goal is to achieve unified communications, unified data, unified access, and unified computing in complex maritime systems, laying the foundation for larger-scale integration and ultimately achieving an unmanned system architecture that is both open and reliable, secure and high-performance. this is where our architectural design work begins. the first is to ensure behavioral consistency; the second is to ensure data consistency; the third is to give full consideration to business agility; the fourth is to decouple functions from data; and the fifth is the independent evolution of management and maintenance.

make full use of the advanced technology of the internet of things to realize the lightweight network proxy architecture, so as to achieve the reliability of the entire system. under the overall mission goal, how to decompose the task, and in the case of task decomposition, how to find the optimal service combination and realize the final task through service combination, this is a relatively technical thing.

overall architecture design of maritime unmanned systems.

in the first part, we said that it is very important, in the second part, we talked about what it is going to solve, and in the third part, we talked about where to start with the system architecture.

first, any product has software and hardware architecture design, second, functional domain design, third, interface specification design, and fourth, architecture integration design.

first, we will give an overall introduction to the software and hardware architecture design, combined with specific products and practices.

first of all, we are working on the sea. the offshore platforms are basically unmanned boats, unmanned ships, drones, etc. the boat side includes structure, power, electricity, autonomous control system, payload equipment, etc. for these things, we make an overall hardware and software access design to achieve heterogeneous device access, device virtualization, and software service scheduling and calling after virtualization. from this perspective, a consistent control interface is exposed to the outside world, making the underlying control process completely transparent; secondly, the upper-level control software and application software of the controller are service-encapsulated to provide overall flexibility of the system.

from the perspective of the physical domain, we need to be able to access relevant payload equipment, which has its own particularity at sea, because at sea, it is relatively clear that the general platform + mission payload is adopted, and an app is added to achieve specific tasks. from this perspective, it is inconsistent with the current platform role and mission singular design principle, which provides a basis for the subsequent capability resourceization.

the second control domain is to abstract and virtualize various types of resources; the second is to be able to independently develop task requirements in a componentized manner, realize task packages for typical scenarios, and be as platform-independent as possible during the scenario development process. this can solve the problem mentioned earlier: whether your product is adaptable to new scenario requirements in the future. it can be adapted, but this adaptation must be done in advance during system design.

the overall plan includes lightweighting of the above-mentioned process application services, networking of the entire intermediate service agent, and abstraction of algorithm loads of all devices. there is related work involved.

what we have achieved so far is that basically all the loads involved at sea, whether it is radar, navigation radar, lidar, main engine, power, etc., we are currently working on no less than twenty kinds of loads, and have achieved overall access and modular replacement and charging.

from the functional domain perspective, there must be service scenarios and task requirements in the architecture design. in order to facilitate the top-level integration and use at the system level in the future, it is necessary to define the general functions and the scope of the general functions. only in this way can we better manage the series of algorithm core technology revolutions brought about by the relevant ecology and formats. from the perspective of unmanned equipment, from the perspective of functional domains, there are seven main functional domains: detection perception domain, self-perception domain, situational awareness domain, autonomous decision-making domain, navigation control domain, mission control domain and operation and maintenance support domain.

there are some specific examples here. the detection and perception domain at sea is mainly sonar, optics, radar, etc. i need to abstract specific payload equipment. after the abstraction, all related payload equipment will be service-oriented to provide service support for sensor management services, environmental perception services, and target attribute perception services, thereby serving the upper-level applications and making the underlying things transparent.

the third interface specification design is to integrate the entire system. we define the hardware and software foundations from the bottom; define the functional scope and functional modules from the middle; and define how to interact with each other, including domain-type information integration mechanisms, including publishing positioning, request response, etc.; and define mutual information between platforms, internal common mechanisms and external common mechanisms between platforms. in addition, we define services and integrate services based on service boundaries. finally, we need a core protocol to manage various integrated services in a unified manner, including virtual equipment, address-independent process scheduling, etc.

as mentioned earlier, the control of the non-registered status of unmanned platforms depends, in a sense, on its degree of openness. if there is a relevant degree of openness in the initial definition and address-independent calls have certain permissions, the flight and status can be defined. this is the specification for virtualized access.

integrated architecture design for maritime unmanned systems.

unmanned equipment is not just an unmanned platform, nor is it that an unmanned platform plus all payload equipment can complete related work. new qualities and new domains bring many new production factors, models, algorithms, data, etc. in addition to the equipment itself, there is a series of subsequent testing, evolution, algorithm warehouses, etc. these are larger resources and require more investment and actual use. it is impossible for us to define an aircraft. for example, it is very difficult to actually go through actual flight certification, but on ships, especially in the field of ships, a large number of foreign classification societies are taking the path of virtual certification, and some work is certified through virtual equipment. the number of airworthiness will be greatly reduced. this is also a feature of the development of unmanned equipment. everyone didn't pay much attention to it, but it suddenly appeared on the battlefield. this is the overall deployment situation.

another aspect is the logical level of data flow relationship and data management, and then how we use the data in related applications and how we realize the feedback of the data to our actual system.

china shipbuilding industry corporation has also developed its own products in terms of architecture, including the xuanlong unmanned boat operating system, and a series of supporting simulation testing and data collection and analysis software, as well as various equipment competitions, products, cluster applications, and a large number of experiments.

future outlook.

in general, our military departments have this requirement, based on national production, suitable for more unmanned platforms, and supporting the overall design of large-scale, distributed maritime unmanned clusters.

the second functional design must be developed in the direction of intelligence;

the third load access should be able to support more service-level service migration and calls.

from my work perspective, there is a lot to learn from the current development status of the low-altitude economy and the development status of military unmanned equipment.

thank you for your listening. this is the end of my report. thank you.