Unmanned aerial vehicle cluster cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics

Abstract

The invention relates to a low-altitude logistics-oriented unmanned aerial vehicle cluster cooperative positioning and intelligent obstacle avoidance system, which relates to the technical field of low-altitude economy and unmanned aerial vehicle cluster logistics, and comprises a channel geometry auditing module, a sensing check identification module, a positioning reference construction module, a time sequence control scheduling module and a beam error suppression control module, a cross-time-scale channel and geometry joint auditing baseline is established, a signal propagation profile is constructed based on urban three-dimensional building appearance data and historical telemetry data, spatial and temporal distribution characteristics of a reflection source are inversed, and an energy turn-back trajectory density map is generated and used for providing dynamic positioning analysis reference. According to the method, an air-ground integrated cooperative sensing and intelligent regulation and control mechanism is constructed, turn-back interference is dynamically eliminated through channel auditing, sensing checking, positioning reference, time sequence scheduling and beam control, and the positioning precision, the obstacle avoidance capability and the flight stability of an unmanned aerial vehicle cluster in a complex low-altitude environment are improved.

Description

Technical Field

[0001] This invention relates to the fields of low-altitude economy and drone swarm logistics, specifically to a drone swarm collaborative positioning and intelligent obstacle avoidance system for low-altitude logistics. Background Technology

[0002] The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics is a comprehensive solution to the problems of inaccurate positioning, delayed obstacle avoidance, and path congestion faced by low-altitude aircraft (including multi-rotor UAVs) performing logistics delivery tasks in complex urban environments. The system's core architecture is based on "cloud-based global scheduling, UAV swarm cooperative perception, and ground-based aircraft-assisted escort." By deploying ultra-wideband (UWB) anchor points at key urban nodes and combining them with visual perception, it achieves high-precision cooperative positioning between UAVs and other low-altitude aircraft. Simultaneously, it utilizes cloud-based digital twin technology to reconstruct the three-dimensional low-altitude environment in real time, dynamically modeling buildings, lines, airspace control, and meteorological elements to provide global path planning and conflict prediction for swarm flights. When a potential collision or sudden obstacle is detected, the system can automatically generate cooperative obstacle avoidance commands based on the principle of global optimization, enabling dynamic avoidance and trajectory adjustment among multiple aircraft. In the last-mile delivery stage, ground-based mobile aircraft (such as robot dogs equipped with UWB beacons and lidar) participate in positioning enhancement and obstacle perception, providing drones with centimeter-level precision landing support and low-altitude safety escort, thereby constructing a closed-loop system covering air-to-ground collaborative perception, collaborative decision-making, and collaborative execution, achieving high safety, high reliability, and high efficiency in low-altitude logistics.

[0003] The existing technology has the following shortcomings: In existing technologies, when aircraft perform low-altitude flight missions between high-rise buildings in cities, their wireless communication and positioning links are often subject to dual interference from obstruction and reflection due to the complex signal propagation environment. This easily leads to multipath propagation phenomena between the metal structures or glass curtain walls of buildings. In this situation, the positioning signal, which should propagate along a straight path, is frequently deflected, resulting in short-term energy superposition and forming a so-called deflection blind zone. Within this blind zone, the aircraft receiver may mistakenly identify the peak energy of the reflected signal as an echo feedback from a real target, causing the control system to misjudge the density of surrounding obstacles and trigger incorrect obstacle avoidance strategies. This phenomenon is particularly pronounced in densely populated high-rise areas. Once the direction of energy superposition deviates from the actual spatial distribution, the aircraft control system may mistakenly trigger emergency deceleration, yaw escape, or other protective actions, disrupting the synchronization of flight paths among the cluster and even causing path crossings, mid-air collisions, or local airspace disorder. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by proposing a drone swarm collaborative positioning and intelligent obstacle avoidance system for low-altitude logistics, thereby solving the problems mentioned in the background.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: a UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics, including a channel geometry audit module, a perception verification and identification module, a positioning reference construction module, a timing control and scheduling module, and a beam error suppression control module: The channel geometry audit module establishes a cross-timescale channel and geometry joint audit baseline, constructs a signal propagation profile based on urban 3D building shape data and historical telemetry data, inverts the spatiotemporal distribution characteristics of the reflection source, and generates an energy return trajectory density map to provide a reference for dynamic positioning analysis. The perception verification and identification module performs time consistency verification of forward-looking perception and side-looking perception, identifies the authenticity of energy peak clusters based on the energy return trajectory density map, extracts misjudged energy areas, and marks the distribution boundaries of false positioning anchor points; The positioning reference construction module generates a self-consistent positioning reference surface based on the distribution boundary of false positioning anchor points, reconstructs the airspace measurement data and ground ranging data in a unified manner, constructs a direct path probability field, eliminates return energy clusters, and forms a continuously updated positioning profile. The timing control and scheduling module constructs a hierarchical timing control strategy based on the continuously updated positioning profile, generates a delay tolerance threshold group according to the reflection density region, calculates the lateral yield curve and time window rearrangement scheme, and outputs the speed, altitude and attitude adjustment sequence for each aircraft. The beam error suppression control module performs dynamic control based on the speed, altitude and attitude adjustment sequence, triggers the dual-polarized beam rotation scanning mechanism, drives the self-consistent positioning reference surface to generate a slip-following effect, unfolds the blanking window in the edge region of the reflected energy, and introduces a time slot misalignment traction process to perform phase cancellation on the transient superposition of reflected energy, thereby realizing cooperative flight closed-loop control in low-altitude environment.

[0006] Preferably, the steps for generating the energy return trajectory density map are as follows: Construct a three-dimensional building geometry model that includes the building outline, facade materials, and surface normal distribution, and register it with historical telemetry data; Based on the location and reflection characteristics of the building surface, signal paths are enumerated, abnormal signal samples that do not match the actual measurements are identified, and the path reflection coefficient is estimated. Cluster analysis of the reflection paths is performed and projected onto a 3D model to generate a return trajectory density map and overlay building facade reflection hotspots; The turnaround trajectory density map is used for dynamic positioning analysis of flight missions, positioning strategy switching logic is set, and the signal model is updated in real time.

[0007] Preferably, the process for calibrating the distribution boundary of false positioning anchor points is as follows: Spatially and temporally register the energy return trajectory density map with the forward visual image and lateral laser point cloud data of the aircraft to extract high reflectivity areas; Perform time-series consistency verification on the registered multi-view perception data, identify abnormal energy peak clusters, and perform path backtracking in conjunction with reflection path data; Extract signal response features within the misjudged area, perform three-dimensional cluster analysis, and determine the spatial distribution boundary of false positioning anchor points; The distribution boundaries of false positioning anchor points are incorporated into the navigation and positioning strategy, the positioning data weights are dynamically adjusted, and the anchor point boundaries are corrected in real time.

[0008] Preferably, the spatial distribution boundary of the false positioning anchor points is jointly determined by the synchronous sensing data of multiple aircraft, and the boundary position is continuously corrected during flight based on the real-time changes in reflected energy to ensure the stability and spatial consistency of the positioning reference data.

[0009] Preferably, the steps for generating the positioning contour are as follows: An initial positioning reference surface is constructed based on the spatial distribution boundary of false positioning anchor points, and high-confidence flight segments are calibrated by integrating aircraft laser point cloud and communication sensing data. Unify ground ranging data and spatial measurement data into a three-dimensional coordinate system, and perform vertical constraints and bidirectional correction of the spatial reference surface; A direct path probability field is constructed based on spatial confidence weights, and spatial nodes that intersect with the reflection path are eliminated to form a continuous path channel; The path channel structure is solidified into a continuously updated positioning profile, and the path error is dynamically corrected by combining real-time and historical data to achieve positioning correction.

[0010] Preferably, when constructing the probability field of the direct path, the confidence assessment of the path nodes is based on point cloud density, ranging consistency and path penetration stability, and a continuous spatial profile is formed by eliminating reflection paths to eliminate the influence of misjudgment.

[0011] Preferably, the steps for generating the velocity, altitude, and attitude adjustment sequences are as follows: Based on the continuously updated positioning profile, the flight airspace reflection density level area is divided, delay tolerance threshold group is generated and associated with response delay range and track offset parameter; Based on the positioning profile and the predicted flight path of the aircraft, a lateral yield curve is generated and a time window rearrangement scheme is constructed to plan the passage sequence of the aircraft in the high-reflection area. The lateral yield curve and time window rearrangement scheme is converted into a sequence of speed, altitude and attitude adjustments, and the timing of execution and adjustment duration are specified. During the execution of the adjustment sequence, the aircraft dynamically compares the actual trajectory with the predicted path, and fine-tunes the adjustment sequence and corrects the path error model in real time based on the offset magnitude.

[0012] Preferably, the lateral yield curve is generated by combining the aircraft's heading difference and spatial interval to plan the offset distance, and a recovery channel is set to guide the aircraft to return to the original track after passing through the high-reflection area.

[0013] Preferably, dynamic control is performed based on velocity, altitude, and attitude adjustment sequences to trigger dual-polarized beam rotation scanning, drive the self-consistent positioning reference plane to generate slip following, expand the blanking window in the edge region of the reflected energy, and introduce time slot misalignment traction to perform phase cancellation steps as follows: Attitude control is executed based on the speed, altitude, and attitude adjustment sequence to drive the dual-polarized beam structure to rotate and scan, and a sliding reference surface is constructed to achieve sensing direction adjustment; Based on the intersection of the slip reference surface and the reflection density boundary, a spatial blanking window is established and a lower limit of the received intensity and a time gate window are set to remove reflected energy. A time-traction strategy is introduced and a staggered time slot is set for the aircraft's sensing trigger point. Different beam scanning angles are bound in the spatial direction and phase cancellation operation is applied. By combining attitude adjustment status, beam response changes and signal confidence assessment results, the intensity and timing of control command execution are continuously dynamically fine-tuned and closed-loop updates are completed.

[0014] Preferably, the slip reference surface generated during beam scanning continuously updates its normal direction and is coupled with the flight path in real time, so that the blanking window only expands outside the reflection energy boundary, and phase cancellation is performed on abnormal direction signals by binding timestamps.

[0015] The beneficial effects of this invention are as follows: This invention establishes a refined cognitive foundation for complex propagation environments through channel geometry auditing, accurately identifies misjudged areas and marks false anchor points using perception verification methods, dynamically eliminates turnaround path interference using a positioning reference structure that integrates airspace and ground data, and simultaneously achieves path misalignment optimization and attitude coordination adjustment through a hierarchical timing scheduling mechanism. Finally, combined with dual-polarized beam scanning and time-slot misalignment traction strategies, it effectively suppresses control mis-triggering caused by transient energy interference. Overall, this scheme constructs a complete mechanism for integrated air-ground collaborative perception, intelligent control, and closed-loop obstacle avoidance in dynamic urban low-altitude environments, improving the flight stability, safety, and mission execution efficiency of UAV swarms in complex airspaces. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the module of the UAV swarm collaborative positioning and intelligent obstacle avoidance system for low-altitude logistics according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0019] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0020] This invention provides, for example Figure 1 The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics shown is characterized by including a channel geometry audit module, a perception verification and identification module, a positioning reference construction module, a timing control and scheduling module, and a beam error suppression control module. The channel geometry audit module establishes a cross-timescale channel and geometry joint audit baseline, constructs a signal propagation profile based on urban 3D building shape data and historical telemetry data, inverts the spatiotemporal distribution characteristics of the reflection source, and generates an energy return trajectory density map to provide a reference for dynamic positioning analysis. When drone swarms perform low-altitude logistics tasks in urban areas, the dense and diverse nature of buildings makes signal propagation susceptible to interference from factors such as obstruction and reflection, thus affecting positioning accuracy. To achieve high-precision positioning in such complex environments, it is necessary to construct a propagation trajectory recognition method with spatiotemporal resolution capabilities to deduce the dynamic distribution patterns of signal reflection sources in urban space, providing an accurate environmental perception foundation for subsequent error correction and path optimization. The specific steps are as follows: Based on high-precision 3D building shape data of urban areas, a geometric model of the flight mission airspace is constructed. This model should cover the outlines, facade materials, surface normal distribution, and precise spatial distribution of visible reflective surfaces of all key buildings within the target flight area. The 3D data must originate from a structured model jointly generated by lidar scanning and aerial photogrammetry, ensuring sufficient geometric accuracy and surface material differentiation capabilities. Based on this, and combined with the trajectory information of the preset flight path, key propagation analysis areas are delineated. Telemetry data collected during past flight missions within the same area is then utilized. This data should include parameters such as the timestamp, signal strength, propagation delay, and angle of arrival of the positioning signals received by the aircraft at specific times and spatial locations. The aforementioned building model and telemetry data are rigorously registered spatially to ensure a clear geometric correspondence between signal measurements and reflection sources, laying an accurate foundation for subsequent propagation profile construction.

[0021] After registering the building's geometric model with telemetry data, signal path reconstruction analysis is performed on the registration results. The specific method involves starting from the theoretical straight-line path between the aircraft and the signal source, and combining the location, angle, and reflection characteristics of the building surface to enumerate and construct all possible single or multiple reflection paths. Each path should possess unique time delay characteristics and spatial reflection point coordinates. By comparing the timestamp of the telemetry signal with the propagation delay of possible paths, it is determined whether the signal originates from a direct path or a reflection path. In the comparative analysis, anomalous signal samples with abnormally amplified signal strength, delayed arrival time, and arrival angle deviating from theoretical expectations are identified and marked as possible echoes from return paths. Simultaneously, the electromagnetic reflection characteristics of the building's exterior surface materials, such as the strong reflectivity of glass curtain walls or the specular reflectivity of metal surfaces, are used to estimate the path reflection coefficient, further improving the accuracy of reflection path identification. The core objective of this stage is to lock down the return trajectories of multipath signals in space and their formation causes through dual constraints of time delay and angle.

[0022] Based on path inversion, all identified reflection paths are clustered and projected onto a 3D geometric model to generate a retracing trajectory density distribution map with spatial location as the main axis and energy density as the metric. This density map needs to be layered and labeled with a time dimension to reflect the changing trends and concentration of reflected energy in different areas over different time periods. Each high-density area represents a high-risk area with frequent reflection source activity and significant energy overlap. To enhance the map's adaptability, temporal reflection hotspot markers of building facades are overlaid on the density map. These hotspots are extracted from long-term telemetry data and reflect that certain building surfaces are more likely to become the dominant source of reflected echoes under specific time periods or meteorological conditions. Furthermore, the density map is reconstructed for continuity in the spatial dimension, enabling it to not only express discrete reflection point information but also present the path coherence and directional trends of the retracing trajectories. The final form of this map is a spatiotemporal reflection profile highly coupled with the 3D environment, providing a reference for predicting the retracing energy corresponding to the flight path of each aircraft.

[0023] After generating the turnaround trajectory density distribution map, it is applied to the dynamic positioning analysis of flight missions. First, for each predetermined flight segment, the corresponding spatial turnaround energy distribution information is extracted to determine when and where the aircraft will enter the high-risk turnaround area. Based on this judgment, the system pre-sets positioning strategy switching logic. For example, before entering a high-energy density area, the weight of the current UWB ranging signal is reduced, and the priority of inter-aircraft visual relative positioning and inertial navigation is increased to avoid being misled by abnormal energy peaks. At the same time, to cope with possible changes in building surfaces during long-term flight operations, such as changes in reflection conditions caused by construction, the actual signal data received by the aircraft in the current flight segment should be compared with the predicted values ​​in the density map in real time to dynamically correct the reliability of the current reflection model. If a new high-energy path is found to deviate from the original model, a local update process should be initiated to add the new path to the profile map to ensure that the turnaround density map maintains the accuracy and real-time performance of its evolution over time. Through the above mechanism, when performing low-altitude flight missions, aircraft can dynamically adjust the positioning source weights and strategies based on a high spatiotemporal resolution reflection energy prediction model, significantly improving positioning stability and anti-misjudgment capability in complex urban environments.

[0024] The perception verification and identification module performs time consistency verification of forward-looking perception and side-looking perception, identifies the authenticity of energy peak clusters based on the energy return trajectory density map, extracts misjudged energy areas, and marks the distribution boundaries of false positioning anchor points; To improve the aircraft's ability to recognize reflected signals in complex urban environments, it is necessary to build an energy return trajectory density map and then further combine it with an airborne multi-view sensing device to perform temporal verification and spatial consistency analysis. This will accurately identify the authenticity of reflected energy peak clusters, extract misjudged areas, and clarify the spatial distribution boundaries of false positioning anchor points, providing a data foundation for subsequent positioning elimination and path reconstruction. The specific steps are as follows: After constructing the energy return trajectory density map, it is spatially and temporally registered with multi-source data collected by the aircraft's onboard sensing equipment. During mission execution, the aircraft continuously senses environmental information ahead and to the sides of its flight path using a forward-facing visual camera and a side-facing lidar. The visual camera is responsible for acquiring color image sequences, while the lidar is responsible for acquiring high-density spatial point cloud data. All data are simultaneously recorded with timestamps and pose information and are jointly calibrated at a predetermined frequency, enabling the fusion of the forward-facing image and the corresponding side-facing point cloud at any given time point within a unified coordinate system. Based on the data registration, key time periods along the continuous flight path are extracted and overlaid with the reflection density data of the corresponding time periods from the previously generated energy return trajectory density map. This identifies potential high-reflection areas and key monitoring areas under the current flight condition. The purpose of this operation is to extract spatial regions highly correlated with the reflection density map from the full range of sensing data, achieving focused data processing and improved efficiency.

[0025] Based on spatiotemporal registration, the fused multi-view perception data undergoes temporal consistency verification. This process uses the time axis as a baseline, comparing the front-view image sequence and the side-view point cloud across consecutive time periods to determine whether the same target exhibits similar spatial response characteristics under different sensor perspectives. Specifically, it identifies structural edges and texture distribution in the images, and spatially abrupt regions in the laser point cloud, confirming whether they represent the same physical entity through spatial matching relationships. In successfully matched regions, spatial feature intensity and its temporal variation trend are extracted to form stable perception regions for real obstacles. Regions with significant reflected energy peaks but unstable perception feature matching are identified as candidate regions for abnormal energy clusters. At this point, reflection path data from the return trajectory density map is used again to backtrack the path of this abnormal region. If multiple densely overlapping return paths from different time periods are found, the probability of it being a misjudged energy source is increased. Through this process, the temporal stability of multi-source perception is combined with the path spatial superposition of the return energy model, enabling accurate identification of potentially misjudged regions.

[0026] After initial screening of misjudged areas, spatial boundary information that may constitute false positioning anchors is further extracted from these areas. Specifically, statistical analysis is performed on the strength and arrival time of the positioning signals received by the aircraft within the misjudged areas to identify recurring spatial points that generate high-intensity responses at multiple aircraft locations but cannot be explained by physical obstructions. These points exhibit significant pseudo-target characteristics and are easily misidentified as true positioning references by flight control decisions. Three-dimensional clustering analysis is performed on these spatial points to extract their distribution range, shape characteristics, and density variations, constructing a candidate set of false anchors. For each candidate anchor set, the historical energy path density variation trend recorded by the return trajectory density map is further superimposed to determine whether the anchor's formation has long-term stability. Spatial regions that simultaneously meet the criteria of multiple aircraft misjudgment records, high return energy density, and intersection with multiple reflection paths are identified as false positioning anchors. Finally, a three-dimensional spatial distribution map of false anchors is generated, containing the position coordinates, radius of influence, and potential interference direction of each anchor point on the flight path, providing a clear basis for elimination in the next step of positioning reconstruction.

[0027] Based on the distribution map of false positioning anchors, this information is incorporated into the real-time navigation and positioning strategy during the aircraft's current flight mission. Within each flight time slice, it is first determined whether the aircraft is about to enter the influence range of a false anchor. If it is predicted that the aircraft will pass through this area in the future, the current positioning data weight allocation is dynamically adjusted, reducing the trust level of wireless positioning data while increasing the response priority of visual positioning and inertial navigation. Furthermore, using new data collected in real time by sensing devices in front of and to the sides of the aircraft, the boundaries of false anchors are finely corrected to prevent them from mistakenly hitting real obstacles due to environmental changes. During flight, the aircraft continuously updates the position and influence range of false anchors based on the correlation between its current perception state and the reflection energy map, ensuring that all sources of false signals are promptly identified and isolated during flight decisions. Through the above perception verification and misjudgment extraction methods, the aircraft can effectively avoid positioning misjudgments caused by the superposition of return paths, maintaining the consistency and collaborative accuracy of the flight path.

[0028] The positioning reference construction module generates a self-consistent positioning reference surface based on the distribution boundary of false positioning anchor points, reconstructs the airspace measurement data and ground ranging data in a unified manner, constructs a direct path probability field, eliminates return energy clusters, and forms a continuously updated positioning profile. To ensure continuous and reliable accurate positioning of UAVs in complex low-altitude urban environments, based on the identified distribution boundaries of false positioning anchor points, it is necessary to further construct a three-dimensional reference structure capable of dynamically eliminating the influence of misjudged signals. This will enable high-confidence real-time positioning support, and a unified reconstruction of airspace and ground data will be performed to ultimately form a continuously updated positioning profile. The specific steps are as follows: Based on the three-dimensional spatial distribution boundary of the identified false positioning anchors, the spatial distribution of these misjudged signal sources is used as the initial reference surface for the elimination area. This boundary information includes not only the spatial coordinates of each false anchor, but also multi-dimensional data such as its interference radius, formation path, and superposition time period. On this basis, real-time perception information is extracted from known UAV flight paths, mainly including spatial point cloud data collected by the aircraft via lidar during flight, and relative position and velocity information shared in inter-aircraft communication. In a three-dimensional coordinate system, this spatial measurement data is spatially fused with the false anchor boundary model to determine the flight path segments that traverse the misjudged area, and the confidence level of the positioning data on these segments is rated. Based on this, a continuous spatial positioning reference surface is constructed, using high-confidence flight segments as the basis and excluding low-confidence areas, to support the structural framework for subsequent positioning judgments.

[0029] To further enhance the spatial integrity and data fusion capabilities of this positioning reference structure, ranging data from ground-based auxiliary units are integrated while constructing the aerial reference surface. These ground-based auxiliary units mainly refer to mobile ground platforms deployed near building edges, ground intersections, or UAV landing zones. Their high-frequency ranging devices provide precise distance data from the ground to the aircraft during low-altitude flight. Through synchronous communication between the aircraft and the ground-based ranging units, the ground-to-air ranging information is mapped to a coordinate system consistent with the airspace measurement data, thereby achieving high-precision vertical constraints in three-dimensional space. Combining the forward flight path and ground-based data feedback, the aerial reference surface is bidirectionally corrected. On the one hand, this eliminates potential path offsets caused by false anchor points in space; on the other hand, it enhances the geometric stability and real-time response capability of the reference surface. The reference surface constructed in this way not only possesses lateral trajectory continuity but also provides stable vertical ranging support, providing a spatial foundation for the next step of constructing a path confidence model.

[0030] After constructing a positioning reference surface that incorporates both airborne and ground-based features, the next step is to utilize multi-source data from this surface to establish a probability distribution structure for direct paths. The core objective of this step is to calibrate in space whether each set of ranging data received by the aircraft originates from a direct path. To achieve this, it is necessary to combine the previously identified false anchor point distribution areas, assigning low probability weights to all ranging responses within these areas and high confidence probabilities to areas supported by both ground and airborne data. In three-dimensional space, based on these probability weights, a confidence distribution field is generated for each spatial point. This field uses point cloud density, ranging consistency, and path penetration stability as core evaluation criteria to filter out the ranging path most likely representing the true distance in a complex reflection environment. For spatial nodes intersecting with multiple reflection paths, they are automatically designated as rejection objects in the probability field, forming a continuous reflection rejection surface within this spatial range, thus removing them from the positioning reference surface. Through this confidence filtering and rejection mechanism, the final constructed path distribution surface retains only ranging data with clear evidence of a direct physical path, thereby forming a highly reliable, stable, and continuous path channel data structure.

[0031] After obtaining the direct path probability field and eliminating reflected energy paths, this structure needs to be solidified into a continuously updated positioning profile to support real-time positioning judgment and path correction during flight. This positioning profile not only reflects the effective direct ranging area in the current space but also dynamically adjusts over time. During actual flight, the aircraft continuously collects real-time data streams from airborne vision, lidar, and ground ranging units. The system spatially matches these data streams with the existing positioning profile to determine whether current environmental changes exceed the original reference surface tolerance range. If a new high-energy reflective cluster or ranging abrupt change occurs in a certain flight segment, the area is marked as an area to be updated, triggering a local reference surface reconstruction process. This reconstruction relies not only on real-time data acquisition but also on historical accumulated data to perform trend fitting of channel change trends within the area, avoiding overcorrection caused by a single abnormal sampling. Ultimately, throughout the entire flight mission, the aircraft uses this continuously updated positioning profile to achieve real-time identification and correction of path deviations and effectively avoid positioning misleading in areas with frequent return energy, ensuring high-precision, safe, and stable flight for each aircraft in the cluster.

[0032] The timing control and scheduling module constructs a hierarchical timing control strategy based on the continuously updated positioning profile, generates a delay tolerance threshold group according to the reflection density region, calculates the lateral yield curve and time window rearrangement scheme, and outputs the speed, altitude and attitude adjustment sequence for each aircraft. To achieve coordinated obstacle avoidance and precise flight control of drone swarms in complex urban airspace, a control strategy with regional differences and temporal dynamics needs to be constructed based on a continuously updated positioning profile. Through hierarchical management and a time-based scheduling mechanism, fine-tuning instructions can be output for each drone, enabling orderly passage through areas of reflection interference and overall path optimization. The specific steps are as follows: Based on the constructed and real-time updated positioning profile, the UAV's designated flight airspace is divided into multiple reflection density level zones. This division is based on the overlap of reflection energy density distribution maps and historical ranging instability data. Through spatial gridding, the reflection level of each airspace node is clearly identified. The level division includes at least three categories: low reflection zone, medium reflection zone, and high reflection zone. The high reflection zone often overlaps with areas where building facades are concentrated, glass curtain walls have high reflection intensity, or mobile reflection sources are frequent. Combining the temporal superposition frequency of reflected energy and historical path disturbance data in each level zone, corresponding time response requirements and track tolerance indicators are set. Furthermore, a delay tolerance threshold set is configured for each reflection density level zone. This threshold set includes the allowable range of aircraft response delay, the upper limit of track deviation, and the minimum interval for obstacle avoidance decision triggering. Through this hierarchical mechanism, dynamic response specifications for different zones are provided for subsequent control strategies, enabling control commands to adaptively adjust according to the environmental conditions of the aircraft.

[0033] After determining the delay tolerance thresholds for each region, the path control requirements of the aircraft are proactively predicted. This requires combining the spatial continuity structure within the positioning profile to predict the time point at which the UAV's current trajectory intersects with the high-reflectivity zone, while simultaneously assessing its speed, direction, and spatial separation from adjacent aircraft at that point. For potential path overlap or response interference, a path offset mechanism based on the spatial yield principle is introduced. Specifically, before the aircraft approaches the boundary of the high-reflectivity zone, a lateral yield curve is designed based on the difference between its flight direction and the heading of the expected encountering aircraft. This curve, while maintaining trajectory continuity, deviates from the center trajectory of the current main path by a safe distance and includes a pre-set recovery channel to ensure the aircraft can smoothly return to its original trajectory after passing through the high-risk area. Meanwhile, in response to the potential time clustering phenomenon in high-reflectivity areas, i.e., the situation where multiple aircraft approach the same airspace point at the same time, a time window rearrangement scheme is constructed to perform fine-grained misalignment processing on the entry time of the aircraft. Within the allowable response delay range, the passage time of each aircraft is redistributed, thereby effectively reducing the probability of instantaneous conflict between airspace nodes and maintaining the overall rhythm of cluster operation.

[0034] After generating the lateral yield curve and time window rearrangement scheme, these are further converted into specific flight control commands, forming a sequence of speed, altitude, and attitude adjustments that can be directly adopted by the onboard control actuators of each UAV. This sequence includes the flight speed vector, vertical altitude value, and pitch, roll, and yaw angle adjustments required by the aircraft at each moment, all output in a format consistent with the aircraft's inertial navigation system coordinate system. During generation, the aircraft's current dynamic constraints and control execution response time must be considered simultaneously. For example, for flight segments approaching high-reflection areas, if the attitude adjustment range is large, gradual adjustments must be made in advance within permissible limits to prevent aircraft oscillations or increased energy consumption due to large-scale operations in a short period. Furthermore, the speed adjustment process must also consider the position of the energy rejection surface along the preceding flight path to avoid rapid acceleration or deceleration in areas with low signal stability, thereby improving the continuity and accuracy of positioning data throughout the entire flight segment. Each adjustment sequence not only includes the target value but also its trigger timing and effective time range, used to determine whether to switch or maintain the current state during flight control execution.

[0035] During the execution of the adjustment sequence by the aircraft, the adaptability of control commands is monitored and corrected in real time in conjunction with the continuously updated positioning profile. When the actual trajectory of the aircraft deviates from the predicted trajectory, or when abnormal signal drift occurs within the originally set delay tolerance threshold, the control sequence needs to be dynamically fine-tuned, and its adjustment tasks on subsequent path segments need to be recalculated. This adjustment process should maintain the original time window misalignment rhythm, and only the aircraft's own velocity curve and attitude angle fine-tuning range are reconstructed in detail to ensure that the entire UAV swarm can successfully complete the operation of crossing high-reflection areas according to the established timing sequence and path allocation requirements during the flight mission. At the same time, in order to improve the long-term adaptability of the path control strategy, after each high-reflection area crossing, the aircraft will send back its displacement deviation and control response delay data during the actual control execution process, and combine it with the current positioning profile to iteratively correct the path error model and optimize the control prediction accuracy of subsequent tasks. Through the above-mentioned multi-level, multi-timescale control strategy construction and adjustment mechanism, efficient response to path interference, multi-aircraft convergence and perception instability in complex urban airspace can be achieved, significantly improving the overall collaborative stability of the flight formation and the spatiotemporal consistency of path execution.

[0036] The beam error suppression control module performs dynamic control based on the speed, altitude and attitude adjustment sequence, triggers the dual-polarized beam rotation scanning mechanism, drives the self-consistent positioning reference surface to generate a slip-following effect, expands the blanking window in the edge region of the reflected energy, and introduces a time slot misalignment traction process to perform phase cancellation on the transient superposition of reflected energy, thereby realizing cooperative flight closed-loop control in low-altitude environment. To ensure high-precision coordinated flight of UAVs in complex low-altitude urban environments, it is necessary to guide them to dynamically execute attitude control and beam response mechanisms based on the generated speed, altitude, and attitude adjustment sequences. This involves actively intervening in the spatial reception characteristics of positioning signals, constructing a slip reference surface and a blanking window, and employing time-traction and phase suppression mechanisms to actively mitigate transient disturbances in reflected energy and maintain stable flight closed-loop control. The specific steps are as follows: Based on the output speed, altitude, and attitude adjustment sequences for each aircraft, the aircraft is gradually driven into the target flight segment, maintaining a dynamic response state for control execution. During flight, the adjustment sequences are used not only to control the aircraft's position and attitude but also as a timing reference for triggering sensing actions. When the aircraft enters a high-reflection energy edge region or traverses a reflection-frequent area along its path, a dynamic adjustment program for the sensing angle is initiated at pre-set time points in the sequence. Specifically, the signal receiving array onboard the aircraft employs a dual-polarized beam structure, which supports adjustment of the direction and angle of the main lobe of the receiving beam in both pitch and roll directions. When the adjustment sequence is triggered, the signal receiving array begins executing a beam rotation scanning program based on the current attitude, periodically scanning the target area with electromagnetic receiving direction within a certain angular velocity range. This process does not rely on the physical steering of the aircraft itself but achieves inertial-free beam adjustment through electromagnetic structure switching, thereby enhancing directional perception while maintaining the flight state. During the beam rotation process, a sliding reference surface is gradually established with the current position of the aircraft as the center and the trajectory of the sensing direction change as the envelope. This reference surface adjusts its normal direction and following path in real time so that its response range to the return signal is always in the position with the least interference, providing spatial conditions for subsequent sensing elimination.

[0037] After the slip reference plane is formed and attitude following adjustment is completed, spatial suppression operations based on reflection density boundaries are further implemented. The main objective of this stage is to generate blanking windows for high-risk areas at the edge of the return energy spectrum, shielding signal reception from interference caused by short-term energy mutations. Specifically, based on the spatial intersection of the return trajectory density map and the current slip reference plane, high-reflection energy gradient boundaries are extracted, and multiple spatial profiles are established around these boundaries according to the sensing angle. In each profile, a lower threshold for received intensity is set based on the mapping relationship between beam reception response characteristics and environmental structure. Any reflection response that does not meet this threshold is blanked out during the data acquisition phase. Simultaneously, a time-gating mechanism is introduced at the receiver end. During each beam rotation, a time-gating window consistent with the path prediction model is established, ensuring that the aircraft only receives energy echoes within the set signal window, thus achieving preventative filtering of transient reflections in the time domain. This dual constraint mechanism of spatial blanking and time gating creates a highly selective sensing zone at the edge of the slip reference plane, allowing only signals with high direct-arrival probability characteristics to pass through, effectively isolating external interference.

[0038] Building upon the blanking window formation, to further suppress residual transient superposition energy, a traction strategy based on temporal distribution and spatial frequency misalignment is introduced. This strategy drives the phase of the aircraft's sensing trigger point to be staggered along the time axis, thereby reducing the probability of echo overlap from multiple aircraft in the same area. This operation requires that the entry time of all aircraft in the high-reflection area be fine-tuned through a pre-adjustment sequence, forming sub-second time slot intervals with adjacent aircraft. Spatially, the staggered time triggers correspond to beam scanning angles in different directions, ensuring that even if multiple aircraft enter the same area within similar time periods, the main lobe directions of their receiving beams do not overlap, avoiding the energy superposition peak problem caused by the signal receiver simultaneously receiving echoes from multiple directions on the same spatial path. Furthermore, at the receiving end, the beam direction is bound to a timestamp, and different phase identifiers are used to distinguish the signal source direction, performing phase cancellation processing on any abnormal signals generated in directions outside the preset path within the time window. This suppression mechanism, based on directional differentiation and time axis misalignment, essentially constructs a reverse interference zone on the physical receiving path, actively weakening instantaneous high-energy waveforms in the interference path and improving the reliability of the signal sensing data.

[0039] Based on the comprehensive execution of beam manipulation, reference plane slip, blanking window generation, and phase cancellation strategies, the entire flight control link is closed-loop. This closed loop no longer relies on static control with a fixed path, but instead uses the triple fusion of sensing data, attitude sequence, and space energy state as the control basis to achieve attitude fine-tuning and path updates under dynamic conditions. In the control judgment of the aircraft at each moment, not only the current trajectory and obstacle avoidance requirements are considered, but the changes in the confidence level of the positioning data caused by the changes in its beam response are also calculated in real time, and the intensity and duration of subsequent control commands are adjusted according to the trend of the confidence interval changes. At the same time, the aircraft compares its motion state on the slip reference plane with the actual triggering result of the blanking window to verify whether the current control strategy has formed signal shielding and path stability at the expected position. If the expectation is not met, the beam angle and time slot triggering interval are readjusted to optimize the subsequent execution strategy. This closed-loop process is executed independently by each aircraft, while also completing collaborative verification through inter-aircraft data exchange. This enables the entire flight cluster to maintain a stable relative formation, synchronized signal response, and consistent trajectory avoidance behavior in the complex low-altitude urban space, thereby achieving highly robust collaborative flight closed-loop control capabilities.

[0040] This invention establishes a sophisticated understanding of complex propagation environments through channel geometry auditing, accurately identifies misjudged areas and marks false anchor points using perception verification methods, dynamically eliminates turnaround path interference using a positioning reference structure that integrates airspace and ground data, and simultaneously achieves path misalignment optimization and attitude coordination adjustment through a hierarchical timing scheduling mechanism. Finally, combined with dual-polarized beam scanning and time slot misalignment traction strategies, it effectively suppresses control mis-triggering caused by transient energy interference. Overall, this solution constructs a complete mechanism for integrated air-ground collaborative perception, intelligent control, and closed-loop obstacle avoidance in dynamic urban low-altitude environments, improving the flight stability, safety, and mission execution efficiency of UAV swarms in complex airspaces.

[0041] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0042] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0047] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A drone swarm collaborative positioning and intelligent obstacle avoidance system for low-altitude logistics, characterized in that, It includes a channel geometry audit module, a perception verification and identification module, a positioning reference construction module, a timing control and scheduling module, and a beam suppression control module. The channel geometry audit module establishes a cross-timescale channel and geometry joint audit baseline, constructs a signal propagation profile based on urban 3D building shape data and historical telemetry data, inverts the spatiotemporal distribution characteristics of the reflection source, and generates an energy return trajectory density map to provide a reference for dynamic positioning analysis. The perception verification and identification module performs time consistency verification of forward-looking perception and side-looking perception, identifies the authenticity of energy peak clusters based on the energy return trajectory density map, extracts misjudged energy areas, and marks the distribution boundaries of false positioning anchor points; The positioning reference construction module generates a self-consistent positioning reference surface based on the distribution boundary of false positioning anchor points, reconstructs the airspace measurement data and ground ranging data in a unified manner, constructs a direct path probability field, eliminates return energy clusters, and forms a positioning profile. The timing control and scheduling module constructs a hierarchical timing control strategy based on the positioning contour, generates a delay tolerance threshold group according to the reflection density region, calculates the lateral yield curve and time window rearrangement scheme, and outputs the speed, altitude and attitude adjustment sequence for each aircraft. The beam error suppression control module performs dynamic control based on the speed, altitude and attitude adjustment sequence, triggers the dual-polarized beam rotation scanning mechanism, drives the self-consistent positioning reference surface to generate a slip-following effect, expands the blanking window in the edge region of the reflected energy, and introduces a time slot misalignment traction process to perform phase cancellation on the transient superposition of reflected energy.

2. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 1, characterized in that, The steps for generating the energy return trajectory density map are as follows: Construct a three-dimensional geometric model of the building and register it with historical telemetry data; Based on the location and reflection characteristics of the building surface, signal paths are enumerated, abnormal signal samples that do not match the actual measurements are identified, and the path reflection coefficient is estimated. Cluster analysis of the reflection paths is performed and projected onto a 3D model to generate a return trajectory density map and overlay building facade reflection hotspots; The turnaround trajectory density map is used for dynamic positioning analysis of flight missions, positioning strategy switching logic is set, and the signal model is updated in real time.

3. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 2, characterized in that, The process of calibrating the distribution boundary of false positioning anchor points is as follows: Spatially and temporally register the energy return trajectory density map with the forward visual image and lateral laser point cloud data of the aircraft to extract high reflectivity areas; Perform time-series consistency verification on the registered multi-view perception data, identify abnormal energy peak clusters, and perform path backtracking in conjunction with reflection path data; Extract signal response features within the misjudged area, perform three-dimensional cluster analysis, and determine the spatial distribution boundary of false positioning anchor points; The distribution boundaries of false positioning anchor points are incorporated into the navigation and positioning strategy, the positioning data weights are dynamically adjusted, and the anchor point boundaries are corrected in real time.

4. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 3, characterized in that, The spatial distribution boundary of false positioning anchor points is determined jointly by synchronous sensing data from multiple aircraft, and the boundary position is continuously corrected during flight based on real-time changes in reflected energy to ensure the stability and spatial consistency of the positioning reference data.

5. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 3, characterized in that, The steps for generating the positioning contour are as follows: An initial positioning reference surface is constructed based on the spatial distribution boundary of false positioning anchor points, and high-confidence flight segments are calibrated by integrating aircraft laser point cloud and communication sensing data. Unify ground ranging data and spatial measurement data into a three-dimensional coordinate system, and perform vertical constraints and bidirectional correction of the spatial reference surface; A direct path probability field is constructed based on spatial confidence weights, and spatial nodes that intersect with the reflection path are eliminated to form a continuous path channel; The path channel structure is solidified into a continuously updated positioning profile, and the path error is dynamically corrected by combining real-time and historical data to achieve positioning correction.

6. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 5, characterized in that, When constructing the probability field of direct paths, the confidence assessment of path nodes is based on point cloud density, ranging consistency and path penetration stability, and a continuous spatial profile is formed by eliminating reflection paths to eliminate the influence of misjudgment.

7. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 5, characterized in that, The steps for generating the velocity, altitude, and attitude adjustment sequences are as follows: Based on the continuously updated positioning profile, the flight airspace reflection density level area is divided, delay tolerance threshold group is generated and associated with response delay range and track offset parameter; Based on the positioning profile and the predicted flight path of the aircraft, a lateral yield curve is generated and a time window rearrangement scheme is constructed to plan the passage sequence of the aircraft in the high-reflection area. The lateral yield curve and time window rearrangement scheme is converted into a sequence of speed, altitude and attitude adjustments, and the timing of execution and adjustment duration are specified. During the execution of the adjustment sequence, the aircraft dynamically compares the actual trajectory with the predicted path, and fine-tunes the adjustment sequence and corrects the path error model in real time based on the offset magnitude.

8. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 7, characterized in that, The lateral yield curve is generated by combining the aircraft's heading difference and spatial interval to plan the offset distance, and a recovery channel is set to guide the aircraft back to the original track after passing through the high-reflection area.

9. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 7, characterized in that, Dynamic control is performed based on velocity, altitude, and attitude adjustment sequences to trigger dual-polarized beam rotation scanning, drive the self-consistent positioning reference plane to generate slip following, expand the blanking window in the edge region of the reflected energy, and introduce time slot misalignment traction to perform phase cancellation steps as follows: Attitude control is executed based on the speed, altitude, and attitude adjustment sequence to drive the dual-polarized beam structure to rotate and scan, and a sliding reference surface is constructed to achieve sensing direction adjustment; Based on the intersection of the slip reference surface and the reflection density boundary, a spatial blanking window is established and a lower limit of the received intensity and a time gate window are set to remove reflected energy. A time-traction strategy is introduced and a staggered time slot is set for the aircraft's sensing trigger point. Different beam scanning angles are bound in the spatial direction and phase cancellation operation is applied. By combining attitude adjustment status, beam response changes and signal confidence assessment results, the intensity and timing of control command execution are continuously dynamically fine-tuned and closed-loop updates are completed.

10. The UAV swarm cooperative positioning and intelligent obstacle avoidance system for low-altitude logistics according to claim 9, characterized in that, The slip reference surface generated during beam scanning continuously updates its normal direction and is coupled with the flight path in real time, so that the blanking window only expands outside the reflection energy boundary, and phase cancellation is performed on abnormal direction signals by binding timestamps.

Images