Embodied Intelligence Anchors Fujian’s High-Quality Development in the Low-Altitude Economy

The low-altitude economy is shifting from reliance on data links and centralized dispatch toward autonomous perception, online decision-making, and cooperative control. Embodied intelligence, organized around a closed loop of perception, cognition, decision-making, and execution, offers a new technical foundation for low-altitude operations in complex and open environments. Low-altitude systems built on embodied intelligence for mountain, island, and offshore wind power scenarios remain in a transition stage between concept validation and engineering demonstration. Fujian combines mountains and seas, numerous islands, and abundant offshore wind resources, giving it prominent demand in grid inspection, maritime supervision, island transportation, and mountain emergency response, while also providing an industrial base in new energy batteries, intelligent manufacturing, and digital infrastructure. This article distinguishes the conceptual boundaries between autonomous flight and embodied intelligence, analyzes the internal logic through which embodied intelligence empowers the low-altitude economy, reviews Fujian’s development foundations and practical constraints, and puts forward recommendations covering provincial-level coordination, joint research, scenario opening, standards development, and governance innovation.

The decisive question for the low-altitude economy is no longer simply whether an aircraft can fly, but whether it can reliably complete a mission in a complex environment. As low-altitude applications move from demonstrative scenarios to productive scenarios, the aircraft faces not only the problem of route execution but a composite problem involving environmental perception, target recognition, path adjustment, risk handling, and mission cooperation. Traditional operating models that depend on ground links, remote control, and static rules are increasingly unable to meet requirements in unstructured environments such as mountains, islands, offshore waters, and forest areas. Against this background, embodied intelligence has acquired a clear necessity for introduction.

It should be stated clearly that embodied intelligence is not a general term for deploying some algorithm on an aircraft, nor is it equivalent to ordinary automatic flight control. Autonomous flight mainly addresses attitude control, trajectory tracking, obstacle avoidance and return-to-home, and mission route execution. Embodied intelligence places greater emphasis on the aircraft’s ability to organize perception, cognition, decision-making, and execution into a task-oriented closed loop within a real environment. The value of embodied intelligence does not lie in replacing autonomous flight levels, but in organizing flight control, scene understanding, mission decision-making, and action execution into a closed-loop capability oriented toward real tasks. In other words, autonomous flight answers whether an aircraft can operate safely according to rules, while embodied intelligence answers whether an aircraft can understand a mission amid environmental change and complete the work. The entry of embodied intelligence into the low-altitude economy is currently at a stage of transition from concept validation to engineering demonstration, and the following discussion proceeds on the basis of this maturity judgment.

Fujian is highly representative on this issue. Mountains and seas coexist, and islands are numerous. Low-altitude technologies have genuine demand in grid operation and maintenance, offshore wind power, maritime patrol, mountain emergency response, and island resupply. At the same time, Fujian has a certain foundation in new energy batteries and intelligent manufacturing, which provides conditions for low-altitude systems to develop toward a higher level of autonomy. Therefore, a discussion of high-quality development of the low-altitude economy in Fujian cannot remain at the general level of equipment manufacturing and scenario listing; it should further address embodied intelligence as a capability variable.

1. The Internal Logic Through Which Embodied Intelligence Enters the Low-Altitude Economy

The entry of embodied intelligence into the low-altitude economy is not a conceptual add-on. It is the result of an evolution in the capability structure of low-altitude systems. Four related shifts explain why embodied intelligence has become relevant at this particular moment, and each of them points to the same conclusion: the capacity to perceive, to judge, and to act must increasingly reside on board the aircraft itself.

1.1 Low-Altitude Operations Are Moving From External Control to Endogenous Autonomy

Early low-altitude aircraft mainly relied on ground dispatch, remote control, and preset rules to complete missions. Under conditions of low density and low complexity, this approach was effective. However, under conditions of valley occlusion, strong offshore winds, weak island network coverage, and frequently appearing dynamic obstacles, reliance on external control alone makes it very difficult to guarantee flight safety and mission continuity. The next stage of low-altitude system development can hardly depend solely on stronger links and more rules; it requires improving the aircraft body’s own environmental understanding and autonomous handling capability. The aircraft is no longer merely an execution terminal but an autonomous system capable of continuously perceiving, judging, and correcting behavior in a real environment. An important change has therefore appeared in the low-altitude economy: the operating logic is shifting from external control to endogenous autonomy. Embodied intelligence is the technical expression of that shift, because it locates the closed loop of perception, cognition, decision-making, and execution within the platform rather than in a distant control room.

1.2 The Center of Competition Is Moving From the Airframe to System Capability

The low-altitude economy is not a simple superposition of aircraft manufacturing and operational services. It is a composite system composed of perception, computing, control, energy, and mission organization. As embodied intelligence enters engineering application, the key factor determining the capability of a low-altitude system is increasingly not the performance of a single airframe but underlying capabilities such as multimodal perception, edge computing, autonomous flight control, swarm cooperation, and power systems. This means that the focus of competition in the low-altitude economy is shifting from who can manufacture an aircraft to who can form an autonomous low-altitude system. For Fujian, what truly deserves attention is not only airframe capability but whether complex scenarios, power systems, digital infrastructure, and system intelligence can be integrated. Among these, multimodal perception, edge computing, and intelligent flight control correspond respectively to information acquisition, support for scene understanding, and the action control stage; only when connected with mission decision-making can they constitute a closed loop of embodied intelligence.

1.3 Complex Scenarios Are Not Only Application Endpoints but Also a Precondition for Technological Maturity

In the past, low-altitude technology was more often used to adapt existing technology to scenarios, but embodied intelligence requires technology to be continuously corrected and evolved in real environments. Only when low-altitude aircraft operate continuously in complex missions can effective data feedback and model iteration be formed. Scenarios are therefore not only the application side but also a precondition for technological maturity. Fujian has a clear advantage on this point. Its mountains, coastal waters, islands, and offshore wind power scenarios are not obstacles to low-altitude economic development; on the contrary, they are important experimental conditions for low-altitude autonomous systems. If a low-altitude system can only operate under preset routes and strong manual intervention, its maturity remains close to concept validation. If it can form stable data feedback, model updating, and safe decision-making in real missions, it is closer to engineering demonstration. Fujian has already formed an application-scenario-oriented approach in fields such as digital low-altitude, marine, and transportation development, yet it still faces problems including information silos, inefficient research and development, and dispersed resources.

1.4 Embodied Intelligence Moves the Value Chain Upward

Once embodied intelligence enters low-altitude systems, the value chain of the low-altitude economy no longer stops at the level of complete aircraft and infrastructure. It extends further to key segments such as onboard computing platforms, multimodal sensors, integrated flight control systems, edge deployment capability, and high specific energy power systems. Fujian already has a certain foundation in new energy batteries and has formed a power battery industrial base led by a leading battery enterprise based in Ningde, which provides industrial support for endurance, safety redundancy, and lightweight design of low-altitude aircraft. It should be noted that high specific energy batteries, aviation power, and complex equipment cannot be directly equated; they still need to pass verification of safety, environmental adaptability, and engineering reliability.

2. Conceptual Boundaries Between Autonomous Flight and Embodied Intelligence

A clear conceptual boundary is necessary because the two terms are frequently conflated in public discussion. Autonomous flight is a mature engineering discipline concerned with predictable behavior under defined rules. Embodied intelligence is concerned with task-level understanding and adaptive action in an open environment. The following table summarizes the distinction.

Dimension Autonomous Flight Embodied Intelligence
Core question Can the aircraft operate safely according to rules Can the aircraft understand a mission and complete it amid environmental change
Main functions Attitude control, trajectory tracking, obstacle avoidance and return, mission route execution Task-oriented closed loop of perception, cognition, decision-making, and execution
Relationship to environment Adapts to predefined constraints Continuously perceives, judges, and corrects behavior in a real environment
Dependence on external infrastructure High dependence on ground links and remote control Reduced dependence on continuous links and remote control
Mission understanding Executes predefined routes Combines scene understanding and target recognition
Maturity Relatively mature in operation Transitioning from concept validation to engineering demonstration

The table makes clear that embodied intelligence does not replace autonomous flight. Rather, embodied intelligence builds on stable flight control and adds the capacity to interpret a scene, decide what to do next, and execute that decision reliably. This layered relationship matters for policy design, because investment in embodied intelligence presupposes a sound foundation in conventional autonomy.

3. Fujian’s Practical Foundation for Combining Embodied Intelligence and the Low-Altitude Economy

Fujian’s position in this field rests on four assets. Each is genuine, and each also carries an internal limitation that must be addressed if the asset is to be converted into capability.

3.1 A High-Complexity and Persistent Low-Altitude Mission Environment

The distinguishing feature of Fujian’s low-altitude scenarios is not their quantity but their high complexity and strong persistence. Mountains, islands, offshore wind farms, power grid corridors, and coastal port and shipping scenarios all place higher demands on the perception, judgment, and execution capabilities of low-altitude aircraft. These scenarios are not one-off demonstration settings; they are real mission environments that exist over the long term and can support sustained validation. Viewed through the framework of embodied intelligence, the particularity of Fujian’s scenarios lies in the fact that four stages are under pressure simultaneously: mountain and offshore environments test the stability of perception; identification of defects in power grids and wind turbines tests scene cognition; weak communication and strong wind conditions test online decision-making; and endurance, payload, and wind resistance test execution reliability. This combination of scenarios, in which multiple stages are stressed at the same time, constitutes the realistic foundation for Fujian to carry out engineering demonstration of low-altitude embodied intelligence.

3.2 Energy Operation and Maintenance as the Most Promising Direction for Early Breakthrough

Judging from Fujian’s existing conditions, grid inspection, substation patrol, and offshore wind power operation and maintenance are fields suitable to serve as leading directions for embodied intelligence. These tasks share common characteristics: high frequency, rigid demand, complex environment, and high labor cost. The requirements they place on aircraft naturally include a closed-loop requirement covering perception, recognition, judgment, path adjustment, and action execution. In other words, energy operation and maintenance is not only a low-altitude application scenario but also an engineering scenario that tests whether the closed loop of perception, cognition, decision-making, and execution can be formed. Therefore, what deserves greater attention in the next stage of Fujian’s low-altitude economy is not whether the general scale of inspection should be expanded, but how to improve autonomous working capability in these scenarios.

3.3 A Certain Power System and Manufacturing Foundation

The development of low-altitude systems based on embodied intelligence cannot be separated from power systems and manufacturing capability. Fujian already has a certain foundation in new energy batteries and advanced manufacturing. Relying on its electrification strategy and the leadership of anchor enterprises, Fujian has formed a new energy battery and advanced materials industrial cluster, but it also faces problems such as an incomplete supply chain and technological discontinuities. This indicates that Fujian does not lack industrial support; rather, the existing support has not yet been fully converted into low-altitude autonomous system capability. Converting that support requires validation of safety, environmental adaptability, and engineering reliability, because high specific energy batteries, aviation power, and complex equipment are not interchangeable categories.

3.4 A Certain Foundation for Regional Coordination That Has Not Yet Formed an Efficient Closed Loop

Research on Fujian’s satellite application industry shows that in some emerging industries Fujian has already formed a pattern of dual cores in the north and south, or a pattern of regional division and coordination, in which Fuzhou and Xiamen concentrate innovation resources while Quanzhou, Putian, Ningde, and other places undertake different types of industrial transfer functions. Combined with the needs of low-altitude embodied intelligence, Fuzhou and Xiamen should focus on algorithm research and development, model training, and test evaluation. Ningde, Quanzhou, and Putian should focus on power systems, flight platforms, and industrial supporting functions. Pingtan, together with typical island, coastal, and mountain scenarios, should focus on real-environment validation. The problem is not whether resources exist, but whether these resources form an effective closed loop around the specific direction of low-altitude embodied intelligence.

Region Suggested Function Representative Setting
Fuzhou and Xiamen Algorithm research and development, model training, test and evaluation platforms Concentrated scientific research and digital resources
Ningde, Quanzhou, and Putian Power systems, flight platforms, industrial supporting capacity New energy battery and intelligent manufacturing base
Pingtan and typical island, coastal wind farm, and mountain grid corridor scenarios Real-environment validation Island resupply, offshore wind operation, mountain emergency response

4. Main Constraints on the Development of Embodied Intelligence and the Low-Altitude Economy in Fujian

Five constraints currently limit the conversion of scenario advantage into capability advantage. They are closely interrelated, and progress on any one of them without progress on the others will yield limited results.

4.1 Insufficient End-Side Intelligence and Key Software and Hardware Capability

A low-altitude system based on embodied intelligence is first of all an end-side intelligent system. It requires the aircraft itself to possess strong real-time perception, edge computing, mission understanding, and action control capabilities rather than relying mainly on cloud or ground support. At present, Fujian already has a certain foundation in power systems, but it remains relatively weak in high-end chips, lightweight high-performance sensors, onboard edge computing platforms, and flight control integration under complex environments. Observed through the framework of perception, cognition, decision-making, and execution, Fujian’s current shortcomings are concentrated in perception reliability, scene understanding capability, and end-side real-time decision-making capability.

4.2 The Innovation Chain, Industrial Chain, and Scenario Chain Are Not Yet Truly Connected

Fujian’s scientific research resources, manufacturing resources, and scenario resources are not absent; they simply have not yet formed an efficient closed loop. Research on new energy materials points out that scientific and technological resources are mainly concentrated in Fuzhou and Xiamen, that high-tech industries in other prefecture-level cities are relatively weak, and that the provincial pattern of scientific and technological resources still needs optimization. At the same time, the matching degree between the needs of leading enterprises and provincial supporting supply is insufficient, and a complete ecosystem has not been formed. This problem is even more evident in the field of embodied intelligence and the low-altitude economy: algorithm research institutions lack experience in aviation systems engineering, while equipment manufacturing enterprises lack multimodal perception and model training capability.

4.3 Insufficient Supply of Real Scenario Data and Inadequate Conditions for Model Iteration

The capability of embodied intelligence needs to be formed through continuous interaction in a real environment. Although Fujian has abundant scenarios, low-altitude environmental data, operational data, and test data remain scattered across different departments and units, and data standards and interfaces are not unified. In an industrial form characterized by interwoven chains and multiple participants, information silos, inefficient research and development, and dispersed resources are important factors constraining collaborative innovation. For low-altitude embodied intelligence, this means that Fujian has scenarios but has not yet fully formed a data supply capability matching those scenarios. In particular, it lacks continuous scenario data capable of supporting model training, mission review, and safety assessment.

4.4 Standards and Localized Testing Systems Are Relatively Lagging

In advancing the low-altitude industry, Fujian still faces problems such as weak infrastructure, insufficient testing capability, and missing operational standards. In particular, it lacks local applicable standards for high-humidity and high-salinity environments, wind shear, and battery thermal management, and it also lacks sufficient localized testing platforms. For low-altitude systems based on embodied intelligence, this shortcoming is even more prominent. Without testing methods and evaluation frameworks adapted to complex environments, autonomous decision-making, and multi-aircraft cooperation, enterprises will find it difficult to form rapid validation and low-cost iteration capability.

4.5 Governance Still Leans Toward a Traditional Equipment Logic

Once a low-altitude system based on embodied intelligence enters a real mission environment, it involves issues such as responsibility allocation, data security, exception handling, swarm cooperation management, and refined airspace governance. Traditional equipment supervision logic is oriented more toward deterministic systems, whereas a low-altitude system based on embodied intelligence contains strong environmental interactivity and behavioral dynamics. Fujian therefore needs to advance technological application while simultaneously adjusting its governance approach.

Constraint Affected Stage Consequence
Weak end-side intelligence and key software and hardware Perception, decision-making, execution Continued reliance on cloud or ground support
Innovation, industrial, and scenario chains not connected Whole chain Algorithm teams lack aviation systems engineering experience; equipment firms lack multimodal perception and model training capability
Insufficient real scenario data supply Cognition, decision-making Model iteration conditions remain inadequate
Lagging standards and localized testing systems Validation Difficulty achieving rapid validation and low-cost iteration
Governance oriented toward traditional equipment logic Deployment Responsibility, data security, exception handling, and airspace governance remain unresolved

5. Policy Recommendations for Advancing Embodied Intelligence and the Low-Altitude Economy in Fujian

Five recommendations follow from the analysis above. They are intended to be mutually reinforcing rather than sequential, and each is directed at a specific constraint identified in the previous section.

5.1 Strengthen Provincial-Level Coordination and Form Differentiated Regional Division of Labor

Fujian should not rely on dispersed advancement by individual localities. Instead, it should arrange functions at the provincial level around research and development, manufacturing, validation, and application. It is recommended that Fuzhou and Xiamen, drawing on their scientific research and digital resources, develop platforms for algorithm research and development, model training, and test evaluation. Ningde, Quanzhou, and Putian, drawing on their manufacturing base, should host power systems, flight platforms, and industrial supporting capacity. Pingtan, together with typical islands, coastal wind farms, and mountain grid corridors, should carry out real-environment validation. The idea of differentiated development with efficient linkage, as proposed in research on satellite applications, has direct reference value here.

5.2 Organize Joint Research Around Key Shortcomings

A joint research mechanism involving government, industry, academia, research institutions, and users should be built, focusing on key directions such as end-side lightweight models, multimodal perception, online decision-making in complex environments, integrated design of flight control, perception, and execution, swarm cooperative scheduling, and high specific energy power platforms. Joint research should not remain at the level of laboratory results; it should simultaneously enter the stages of pilot-scale validation and system integration. For Fujian, research projects should be bound to real missions, and the quality of technology should be tested in scenarios such as energy operation and maintenance, offshore wind power, island resupply, and mountain emergency response.

5.3 Build Continuous Validation and Data Opening Mechanisms Based on Typical Scenarios

Fujian’s advantage lies in its complex scenarios. The key to the next stage is not to propose scenario requirements in general terms but to institutionalize scenario development. It is recommended that priority be given to building leading pilot zones for low-altitude embodied intelligence around grid inspection, substation patrol, and offshore wind power operation and maintenance, and that verifiable indicators be formed, including defect identification accuracy, close-approach stability, adaptability to complex weather, endurance redundancy, return safety, and mission review mechanisms. On this basis, the work can gradually be extended to scenarios such as island transportation, maritime patrol, and mountain emergency response, promoting a feedback chain of real mission, data accumulation, model training, and system optimization. The focus of scenario opening is not simply to increase flight activity but to provide a stable environment for technological iteration.

5.4 Give Full Play to the Role of Scientific and Technological Societies and Promote Standards and Evaluation Systems

Low-altitude systems based on embodied intelligence involve the intersection of multiple disciplines, multiple participants, and multiple stages, and standards development cannot rely entirely on a single enterprise. Scientific and technological societies can play a pivotal role in resource aggregation, rule design, platform building, and standards advocacy. At the same time, relevant materials from the standards committee on humanoid robots and embodied intelligence emphasize that standards supply should adhere to agile iteration, whole-chain coordination, and application first. Fujian should promptly promote the development of standards for data interfaces, testing methods, model evaluation, multi-aircraft cooperation, and operation specifications in complex environments, and should establish a testable and reviewable evaluation system around the four stages of perception, cognition, decision-making, and execution.

5.5 Improve Experimental Supervision and Risk Governance in an Inclusive and Prudent Manner

Fujian should explore the establishment of low-altitude intelligent flight experimental zones in Pingtan, in areas with concentrated island scenarios, or in typical mountain areas, implementing an experimental supervision mechanism characterized by limited boundaries, tiered access, whole-process traceability, and dynamic evaluation. The focus of governance is not simply to relax or mechanically tighten rules but to allow rules to adjust gradually as technology evolves, so that technological validation and rule improvement advance in parallel. For low-altitude systems based on embodied intelligence that are still in the transition from concept validation to engineering demonstration, necessary human-machine collaborative takeover and mission review mechanisms should be retained, so that technological experimentation, risk identification, and institutional improvement form a closed loop.

6. Directions of Key Breakthrough

The following table summarizes the directions in which concentrated effort is most needed, together with the capability each direction is expected to unlock. The directions are derived directly from the constraints discussed above and correspond to specific stages of the embodied intelligence loop.

Direction Primary Emphasis Stage of the Closed Loop Most Affected
End-side lightweight models Onboard edge computing capability Decision-making
Multimodal perception Information acquisition and scene understanding Perception and cognition
Online decision-making in complex environments Autonomous judgment under weak communication and strong wind Decision-making
Integrated flight control, perception, and execution Integrated design of control and action Execution
Swarm cooperative scheduling Multi-aircraft mission cooperation Cognition and decision-making
High specific energy power platforms Endurance, safety redundancy, and lightweight design Execution

7. Verifiable Indicators for Leading Pilot Scenarios

A pilot scenario is only meaningful when it produces measurable evidence. The indicators below are proposed as the basis for acceptance in the leading pilot zones, and they are intended to be reviewed continuously rather than applied as a single final test.

Leading Pilot Scenario Verifiable Indicators
Power grid inspection Defect identification accuracy, close-approach stability, adaptability to complex weather, endurance redundancy, return safety, mission review mechanism
Substation patrol Perception reliability, scene understanding, exception handling
Offshore wind power operation and maintenance Perception stability under complex weather, online decision-making quality, execution reliability

8. Conclusion

The entry of embodied intelligence into the low-altitude economy is not a conceptual addition to existing low-altitude technology. It is the result of an evolution in the capability structure of low-altitude systems. Its core significance lies in three shifts: it pushes low-altitude aircraft from being connectable and controllable toward being able to perceive, judge, and execute; it pushes the low-altitude industry from an airframe orientation toward a system capability orientation; and it pushes low-altitude applications from a low-complexity scenario orientation toward sustained mission execution in complex environments.

Fujian has strong practical significance in this direction. Fujian possesses complex scenarios including combined mountains and seas, distributed islands, and offshore wind power, and it also has a certain industrial foundation in new energy batteries, intelligent manufacturing, and digital infrastructure. What truly needs to be resolved is how to organize scenarios, technology, data, standards, and governance into a mechanism that can advance sustainably. The focus of the next stage should shift toward capability building for low-altitude systems based on embodied intelligence. Only by converting the advantage of complex scenarios into an advantage in technological iteration, converting the manufacturing base into an advantage in system capability, and converting institutional space into an advantage in governance innovation can Fujian form a development path with clear local identity.

The wider lesson is that embodied intelligence is not a distant research aspiration for the low-altitude economy but a practical requirement emerging from the environments in which low-altitude missions actually take place. Where links are unreliable, where weather is severe, where terrain blocks line of sight, and where human intervention cannot arrive in time, the value of embodied intelligence becomes concrete. Fujian’s combination of demanding environments and existing industrial capacity places it in a position to test that proposition at scale, provided that coordination, standards, data sharing, and governance evolve alongside the technology itself.

The recommendations set out above are intended to be mutually reinforcing rather than sequential. Provincial coordination establishes the spatial division of labor; joint research addresses the capability shortfalls that no single actor can resolve alone; scenario opening supplies the continuous data that embodied intelligence requires for iteration; standards and evaluation systems make results comparable and reviewable; and inclusive, prudent supervision allows experimentation to proceed without abandoning accountability. Taken together, these elements form the institutional counterpart to the technical closed loop of perception, cognition, decision-making, and execution that defines embodied intelligence.

Scroll to Top