236 results about "Priority queue" patented technology

The invention relates to the technical field of cloud side end computing, and discloses a cloud side end cooperative computing resource dynamic allocation method based on task-aware migration. The method comprises the following steps: acquiring real-time load characteristics and resource demand characteristics of calculation tasks in a cloud side end system, and dividing task priority queues in combination with task type identifiers; extracting historical execution records of tasks at cloud, edges and terminal nodes, constructing a task execution feature library, and generating a resource demand prediction model in combination with real-time load features; analyzing network transmission time delay characteristics of a cloud end and edge nodes, measuring real-time calculation capability fluctuation data of terminal equipment, and establishing an inter-node resource collaboration degree evaluation matrix; generating an initial migration strategy according to the prediction model and the evaluation matrix, monitoring actual resource occupancy deviation of the task, forming a final decision in combination with a node resource state correction strategy, triggering cross-node migration, and synchronously updating the priority queue and the evaluation matrix.

The invention relates to the technical field of intelligent monitoring, in particular to a multi-thread low-power-consumption intelligent monitoring system based on an AI processor. The method has the advantages that aiming at the problems of unbalanced computing power and power consumption, high multi-task processing delay and strong hardware dependence in the prior art, the NPU module of the RK3588 processor is combined with the INT8 quantitative model, so that the power consumption is lower than 10W under the 6TOPS computing power; a dynamic multi-thread scheduling mechanism is designed, parallel processing of more than eight paths of video streams is supported through binding of a priority queue and an NPU core, and end-to-end delay is compressed to be within 200 ms; a zero-copy video stream architecture is constructed, data transfer is eliminated through memory mapping, and preprocessing time consumption is reduced by 90%; an energy efficiency control module is integrated, the NPU voltage frequency is dynamically adjusted according to the load, and the energy efficiency ratio reaches 0.83 TOPS / W; space-time alignment of multi-model reasoning results is realized by adopting a frame ID synchronization technology, and the mismatching rate is lower than 0.1%.

The invention provides a power grid operation analysis method based on a cloud platform, and relates to the technical field of power grid operation analysis, and the method comprises the steps: configuring a semantic event template library, configuring multi-source power grid operation data access and time reference synchronization, configuring an event triggering rule and an event judgment mechanism, and extracting an event fragment data set, and configuring a semantic abstract vector, executing semantic compression, generating a unique analysis task number, registering a task metadata directory, and setting a task priority and priority queue arrangement. By configuring the semantic event template library and realizing the access and time reference synchronization of the multi-source power grid operation data, the unity and timeliness of the event fragment data set can be ensured in a mass data environment, the problem of analysis errors caused by non-uniform data formats and non-uniform time sequences in the prior art is solved, and the analysis accuracy is improved. Therefore, the reliability of power grid operation state identification and analysis results is improved.

The invention discloses a CAN message priority dynamic arbitration method and system based on event triggering, and the method comprises the steps: obtaining a discrete sampling signal of a CAN node of an engineering vehicle and a state register value of a CAN controller, converting the discrete sampling signal into a working condition data sequence, and obtaining a bus load rate through the state register value; obtaining an event state sequence according to the working condition data sequence; all CAN messages to be sent are constructed into a message data record, and the message data record with event semantics and dynamic priority labels is obtained according to the protocol data unit and the event state sequence; and according to the bus load rate and the event state sequence, sending a message data record with event semantics and dynamic priority annotations by adopting a token bucket mechanism through a three-level priority queue linked with a three-state state machine. According to the invention, the security key message in the system can still be transmitted in time under high load, and the bandwidth occupation and delay jitter caused by the low-priority message are avoided at the same time.

The invention belongs to the technical field of cache prefetching optimization of a firmware layer of a computer storage system, and particularly provides a memory control method, a storage device, a medium and equipment.The instantaneous flow is effectively filtered and persistent hotspot data is accurately recognized through a three-layer structure of combining a first-stage hotspot filtering module with a window cache, a view queue and a protection queue; the second-stage continuous access filtering module dynamically judges a continuous access mode by means of mean-variance statistics of an annular buffer area, and accidental continuous triggering of prefetching is avoided; and the third-stage random access judgment module prohibits invalid prefetching and effectively distinguishes access modes. The pre-fetching bandwidth is limited through the token bucket algorithm, the global priority queue of the skip list is combined, it is ensured that the priority of the write task is higher than that of the pre-fetching task, pre-fetching and foreground I / O resources are isolated, delay jitter control over foreground key I / O can be achieved during full-load pre-fetching, and the problem that delay soaring is caused when pre-fetching preempts the resources is solved.

The invention belongs to the technical field of data center network and clock synchronization, and discloses a dynamic time synchronization path optimization method and system based on topology awareness. The invention provides a closed-loop framework of state awareness-path optimization-resource assurance-dynamic adjustment and optimization: in the first stage, network topology, link load and PTP message performance data are periodically acquired through an SDN controller, and a global state view is constructed; in the second stage, the optimal path is dynamically calculated for the PTP synchronous flow through a multi-constraint optimization algorithm with the purpose of minimizing transmission delay jitter, and deterministic transmission is achieved in combination with bandwidth reservation and a priority queue; and meanwhile, a seamless switching mechanism of building first and disassembling second is introduced to adapt to the network state change in real time. On the basis of multiplexing existing hardware, the time synchronization error is remarkably reduced, the transmission stability and the dynamic adaptability are improved, microsecond-level time synchronization can be efficiently achieved, and the method is suitable for high-precision service scenes.

The invention discloses a statistical data downloading method for supply chain finance, which relates to the technical field of finance and comprises an acceptance permission judgment module, a query resource control module and a desensitization packaging evidence storage module. The acceptance permission judgment module is used for uniformly accessing the statistical application, completing identity verification, main body range binding, index range binding, time range binding and area range binding, and generating an acceptance bill and a judgment bill; the query resource control module is used for generating a query plan under the index aperture mapping dictionary, the partition information and the resource quota, binding a concurrent quota, a throttling quota and a priority queue, and outputting an executable instruction stream; and the desensitization packaging evidence storage module is used for executing differential desensitization according to the main body type and the sensitivity level, performing aggregation calculation and deduplication execution, completing result set packaging, digital watermarking and downloading delivery, generating a downloading certificate and writing a certificate abstract and an evidence index into an evidence storage chain. The method has the characteristic of high consistency.

The invention provides a resource loading method and system based on user behavior prediction, and relates to the technical field of data processing, and the method comprises the steps: collecting user behavior data; preprocessing the user behavior data; extracting data features of the preprocessed user behavior data; according to the data features, predicting resource demand weight distribution of the user through a double-layer LSTM based on fusion of an attention mechanism and interest drift perception; calculating the priority of each resource according to the prediction result of the resource demand weight distribution of the user in combination with the importance coefficient, the resource value density and the network adaptation coefficient of each resource; constructing a resource loading priority queue according to the priority of each resource from high to low; and dynamically adjusting a loading strategy of each resource in the resource loading priority queue in combination with a network state, and loading the resources as required. According to the method, the resources can be loaded as required, the resources possibly required by the user are preferentially loaded, and the resource utilization efficiency is improved.

The invention relates to the technical field of network resource management, and discloses an intelligent building weak current equipment multi-protocol compatible method and system, and a storage medium. The method comprises the steps of collecting multi-protocol network data streams, extracting data packet lengths and arrival interval time as classification features, dividing network data packets into three data packet groups of short and small high frequency, medium stability and large flow, and forming a real-time response network sequence through a priority queue scheduling mechanism and preferentially transmitting the short and small high frequency groups, the method comprises the steps of receiving a real-time response sequence, calculating residual bandwidth to adjust a network buffer area, packaging and splitting a large-flow data packet group to generate transmission sub-blocks, merging the real-time response sequence and the transmission sub-blocks through a merging rule, dividing the merged real-time response sequence and the transmission sub-blocks into a plurality of data stream segments, and distributing a transmission path for each data stream segment. And monitoring bandwidth occupation and transmission delay of each transmission path, and dynamically adjusting path distribution of the data stream segments. According to the invention, the utilization efficiency of network resources and the real-time performance of data transmission in a multi-protocol environment are improved.

The invention discloses a translation memory bank retrieval and matching method based on semantic similarity, which relates to the technical field of computer-aided translation, and comprises the following steps: generating an internal edge and a cross-graph edge, calculating a one-time jump success probability, calculating edge and cross-graph edge cost based on cleanliness, correcting by using the one-time jump success probability to obtain an edge weight; generating a weighted graph, constructing an initial priority queue, searching by using a shortest path to obtain a distance matrix, calculating density based on a fuzzy similarity matrix, screening the density to obtain a genealogy center, calculating proximity, and generating candidate translations in combination with the distance matrix; static and dynamic fragment sets are constructed through a dependency syntax and BERT-NER, semantic matching precision and noise robustness are improved, a weighted graph is constructed in combination with cleanliness and transfer factors, shortest path matching is carried out, and the retrieval recall rate and quality of technical translation are improved.

The invention belongs to the related technical field of vehicle charging scheduling, and provides a vehicle charging dynamic scheduling method and system for a charging pile fault in order to solve the problem that a charging device cannot respond in time after a fault occurs, and the method comprises the following steps: monitoring the operation state of each charging pile in the execution of an initial charging scheduling scheme in real time; identifying a target vehicle set influenced by the charging pile fault; calculating an emergency priority score of each target vehicle according to the residual electric quantity and the planned departure time of the target vehicle; constructing a dynamic resource profile of the available charging pile based on the residual available power of the available charging pile, and screening a feasible candidate pile set; and according to the vehicle priority queue, reallocating available charging piles to each target vehicle in sequence by taking the minimum cost as a target, and generating a rescheduling scheme. According to the invention, charging of the vehicle at the valley point of the electricity price can be guaranteed to the greatest extent, and the reliability, economy and adaptive capability of charging are remarkably improved.

A communication method and a related apparatus, which relate to the technical field of communications. The method can be applied to a computing node, and comprises: receiving priority indication information of a computing task to be processed, wherein the priority indication information indicates the processing priority of a computing node for the computing task, and the priority indication information is obtained on the basis of at least one of a task requirement of the computing task and subscription data of a terminal device; and on the basis of the priority indication information, scheduling the computing task to a first position of a first queue or a second queue of the computing node, or directly scheduling the computing task to a third queue, wherein the first queue is one of a plurality of different priority queues of the computing node, the second queue is a queue in the computing node that waits to access a coprocessor resource, and the third queue is a queue in a coprocessor in the computing node. The present application can ensure the quality of service for important users or computing tasks requiring a low latency.

The invention discloses a water environment monitoring and early warning method based on multi-agent collaboration and a block chain, and relates to the technical field of water environment monitoring. The method comprises the following steps: collecting water quality parameters through monitoring equipment deployed in a water area, carrying out data cleaning and feature extraction on the water quality parameters through edge nodes, and executing lightweight anomaly detection; generating a data abstract from the monitoring data after edge processing, and writing the data abstract into a block chain network for credible evidence storage; each edge point serves as an independent agent, shares state data, performs task negotiation and distribution through a game theory algorithm based on a utility function, and performs uplink storage evidence on an intermediate result in a collaboration process; when the monitoring data meets a preset condition, each agent performs collaborative voting, and the intelligent contract automatically verifies a voting consensus and triggers an early warning event of a corresponding level; and according to the triggered early warning level, emergency equipment is automatically dispatched based on the priority queue for response, and an equipment execution result is fed back to a chain to form closed-loop management.

The application belongs to the field of chip shared cache usage method, and relates to a method for sharing cache space among multiple priority queues, which divides the queues according to priorities, allocates different amounts of shared cache space for different priorities, and the allocation principle firstly guarantees the demand of high-priority queues for shared cache space; the corresponding shared cache space is allocated according to the demand of each priority queue for shared cache; the high-priority queue can occupy the idle shared cache space of the low-priority queue, and the occupation is performed according to the principle from low to high priority; a constraint condition is set, when the constraint condition is met, the low-priority queue can occupy the shared cache space of the high-priority queue, and the amount of shared cache space occupied by the low-priority queue is constrained. The application further improves the utilization of shared cache space on the basis of guaranteeing the cache space of high-priority queues.

The invention discloses a video code rate dynamic allocation compression method based on content complexity prediction, which comprises the following steps of: S1, acquiring a video frame sequence, and preprocessing; s2, inputting the frame-level feature tensor into a gated residual convolutional network, and outputting a complexity prediction sequence; s3, constructing a frame priority queue, and calculating the complexity jump amplitude between adjacent frames; s4, a compression area is divided, and a corresponding code rate resource scale factor is allocated; s5, configuring a reference frame structure, a prediction interval and an initial quantization step size for each compression region, and calculating a region target bit number; s6, distributing regional bits to each frame in a compression coding process, and dynamically adjusting a frame-level quantization parameter and an entropy coding strategy; and S7, after compression is completed, reversely updating convolution prediction network parameters through bit distribution errors. According to the invention, fine code rate dynamic allocation and adaptive compression control based on content complexity are realized, and the video compression quality and bit utilization efficiency are effectively improved.

The invention relates to the technical field of voice recognition control, in particular to a Bluetooth sound effect voice interaction method and system, and the method comprises the following steps: completing the energy change analysis and grading recognition based on a voice signal collected by a Bluetooth sound, analyzing channel characteristics in combination with recognition characteristics, summarizing a stable trend, adjusting a priority queue of a sound effect control instruction, and completing the voice interaction of the sound effect of the Bluetooth sound. And matching data packets with different structures and optimizing a distribution process to obtain a delay distribution index. According to the method, voice recognition results are identified in a grading manner, sound effect control instruction priorities are distinguished and sorted in combination with channel stability trend parameters, transmission requirements and feedback responses of different queue instructions are distinguished in a hierarchical data packet structure, task scheduling, the data packet structure and time delay feedback form closed-loop dynamic adjustment, and the control efficiency is improved. Reasonable distribution and feedback management in an instruction transmission process are improved, task response timeliness is continuously optimized, and continuous stability and execution consistency under multi-round sound effect voice interaction are enhanced.

The invention belongs to the technical field of task scheduling of a storage system, and discloses a user-mode-oriented storage task self-adaptive scheduling method and system, and a more efficient task processing framework is realized by replacing an original kernel thread-based storage task scheduling scheme with a user-mode coroutine scheduling mechanism. The multi-level priority queues are directly managed in task scheduling, and CPU core binding configuration is carried out, so that the switching delay of a user mode / kernel mode is reduced, and the task execution controllability is improved. A dual-threshold self-adaptive scheduling strategy switching technology is introduced, and a greedy scheduler and a reinforcement learning scheduler are combined, so that the problem of adaptability of a static scheduling strategy in a dynamic load environment is relieved, performance loss caused by scheduling strategy solidification is avoided, and the self-adaptive capability and service quality guarantee of a system under a complex load are improved. Compared with an existing task scheduling method based on a kernel thread, the task scheduling method has the advantages that the overall performance and the resource utilization rate of a storage system can be remarkably improved while the task execution efficiency and the SLA requirement are guaranteed.

The invention, which relates to the technical field of interactive script generation, discloses a script-centered interaction-driven scene generation system comprising the following modules: a script management module, a script arrangement module, a dependency propagation module, a thesaurus interaction cooperation module and an execution output module. Creating a basic data structure of the script, distributing script attributes, and managing script elements, constraints and logics; identifying scene nodes and events, constructing three-layer dependency and weighting, and generating a reverse causal chain through recursive dependency and depth-first search; by maintaining a dependency table, event-driven updating and dirty node marking for each node, efficient dependency propagation and low-delay updating are realized through a priority queue and differential generation; dividing three levels to cooperatively realize dynamic script generation and interaction, and realizing script control based on gating residual connection, a reverse causal chain and a time attention mechanism; and constructing an interactive execution layer, integrating a real-time feedback mechanism, automatically generating a script, and outputting and displaying an executable format.

The invention relates to the technical field of space-time big data processing, and discloses a Web-end-oriented space-time big data efficient rendering optimization method and system, a storage medium and electronic equipment, and the method carries out unified and intelligent management on the whole rendering process by introducing a central rendering task scheduler. Space-time big data fragments are split into fine-grained vector tile rendering tasks, a dynamic rendering priority score is calculated for each task, and the vector tile rendering tasks with high priorities are sequentially taken out from the head of a dynamic priority queue according to a current available rendering resource threshold value of a Web end. The method comprises the following steps: acquiring a plurality of rendering tasks, distributing the rendering tasks to a plurality of parallel execution rendering task execution units for real-time rendering processing, outputting a final spatio-temporal data visual image, and dynamically adjusting a task execution sequence and resource allocation in combination with real-time system monitoring, thereby preferentially ensuring that contents most concerned by a user are quickly rendered under limited Web end resources, and improving the user experience. And an efficient and smooth visual effect is realized.

This invention relates to the field of gene data analysis technology, and discloses a gene data analysis method and system based on GPU acceleration. The method involves: sorting the resources of each GPU card in a GPU cluster to obtain a first resource sorting queue and calculating the target GPU memory usage of sub-regions in each genome; simultaneously constructing a subtask priority queue; matching the GPU cards for each subtask in the subtask priority queue and constructing a subtask binding list for each GPU card; constructing a Pileup tensor for each bound subtask on each GPU card, performing convolutional classification on the Pileup tensor, and outputting the probabilities of different genotypes. This invention overcomes the technical problem of existing static allocation strategies being completely unaware of the real-time load status of GPU cards, ensuring that subtask allocation decisions are always based on the latest resource status of the cluster, systematically eliminating load skew problems from the scheduling layer.

The invention discloses a dynamically-graded satellite TBTP efficient broadcasting method and a satellite communication system, and the method comprises the steps: obtaining the dynamic forward signal-to-noise ratios of all online terminals in a current network at a transmitting side, and carrying out the dynamic grading, thereby obtaining a multi-stage parallel forward channel broadcasting TBTP sub-table; the scheduling time slot entries are classified into corresponding TBTP sub-tables according to the SNR of the terminal, the TBTP sub-tables are sent to a gateway according to an ascending sequence of the SNR, and the lowest SNR of the grade is carried in each sub-table to serve as a grading SNR identifier; the gateway regards the received sorted TBTP sub-tables as multiple paths of data streams with the same scheduling priority and incorporates the multiple paths of data streams into a TBTP broadcast priority queue; and carrying out physical frame assembly and modulation according to the TBTP broadcast priority queue, and sending the TBTP broadcast priority queue to a receiving side to parse the parsable TBTP signaling so as to restore the parsable TBTP signaling to obtain a complete resource allocation view for reverse transmission. According to the dynamic difference of the forward signal-to-noise ratios of the terminals, differentiated efficient broadcasting of the TBTP is realized, so that the overall utilization efficiency of forward link resources is remarkably improved on the premise of ensuring reliable reception of all the terminals.

Various implementations directed to price time priority queue for a multi-dimension map tile device repository are provided. In one implementation, a method may include receiving origin location data and destination location data. The method may also include generating data networks based on the optimized origin location data and the destination location data. The method may further include determining data hubs along the transmission or transit route and network, where the virtual hubs include a first virtual hub based on the origin location data and a second virtual hub based on the destination location data. The method may additionally include receiving IoT device data for the geolocation exchange units. In addition, the method may include receiving market depth data for a geolocation exchange for the geolocation exchange units based on the multi-dimension map tile repository nodal sequences.

The present application relates to the technical field of data transmission, and specifically provides a data transmission method and system for solid waste sorting equipment, a terminal and a storage medium, comprising: adding basic classification marks and dynamic priority marks to data packets according to the function types of sorting data, and dividing the data into synchronous control flow, real-time event flow, process awareness flow and background management flow; a data source end acquires link state information before sending, and adaptively adjusts the dynamic priority of the real-time event flow; a network switch performs mixed scheduling transmission on the data packets based on time-aware shaping, credit shaping and preemption mechanism, wherein the preemption of low-priority queues is only allowed when the data packets of the highest-priority sub-queue carry the highest dynamic priority mark. Through the cooperative mechanism of basic classification and dynamic priority, the present application realizes fine classification transmission of heterogeneous data flow, solves the problem that key instructions are easily affected by network congestion, and effectively reduces transmission delay and jitter.

The application discloses a kind of based on flow perception intelligent network interface card optimization system and method, it belongs to computer network technical field.The optimization system of the application includes flow perception module, flow prediction module, resource scheduling module, priority queue module and load balancing module.The application introduces resource monitoring, flow perception, resource prediction and dynamic scheduling module, real-time monitoring network flow and predicting its change trend, according to flow mode dynamically adjusts resource allocation.The application effectively solves the problem of resource contention, through priority ordering and load balancing, ensure the stable operation of high priority task, improve the throughput and stability of system, reduce resource waste.The application is suitable for resource management of intelligent network interface card in data center and cloud computing environment, improves network processing efficiency and resource utilization, guarantees service quality, reduces resource waste, and enhances the adaptive capacity of system to dynamic load and complex flow mode.

The invention discloses image processing equipment based on a computer control technology. The image processing equipment comprises a server control and management module which realizes automatic fault switching and load balancing through a cluster formed by a plurality of servers; the image processing module is responsible for receiving an image processing task from the server and executing corresponding processing operation; the image acquisition and transmission module is used for capturing to-be-processed image data from a terminal or other sources through image acquisition equipment; the communication network management module is responsible for managing a communication network; according to the method, a resource reservation mechanism is introduced, enough processing resources are reserved for the burst task according to historical data and a prediction model, rapid response and effective processing can be achieved when the burst task is faced, performance reduction or faults caused by insufficient resources are avoided, a job priority mechanism is established, jobs are ranked, and through priority queue management, the task processing efficiency is improved. And the key operation can preferentially obtain processing resources.

The invention discloses a multi-operator coarse-grained parallel scheduling method based on a heterogeneous data stream architecture, which belongs to the field of runtime systems, and comprises the following steps of: dividing a computing load of an operator with in-degree of 0 in a computing graph to be executed into a plurality of sub-tasks Task and packaging the sub-tasks into independent Task objects; constructing four priority queues which are respectively used for storing Task objects which have different preferences on four functional parts in the processing unit, putting index Task objects of operators into the corresponding priority queues according to the preferences of the operators on the functional parts and the number of instructions of the operators, and sorting the index Task objects, k (K < = 4) Tasks with complementary resource requirements are selected from the four priority queues to form a new composite operator, and the new composite operator is executed by the processing unit array; according to the method, the new operator is generated by identifying the preference of the operator function part and dynamically combining the Task, so that the utilization rate of the internal function part of the PE can be fully improved, and the operation efficiency of the whole data flow program is accelerated.

The present application relates to the technical field of information security, and aims at the problems that password guessing based on probabilistic context-free grammar is difficult to identify multi-semantic patterns and the semantic guidance strength is uncontrollable, and proposes a password guessing method based on multi-semantic fusion of probabilistic context-free grammar: enumerating sub-strings in the training password, identifying semantic segments according to simple, date, vocabulary and name patterns and prioritizing disambiguation, dynamically planning segmentation according to the principle of maximum semantic coverage and least semantic segments, setting semantic enhancement parameters β for each semantic pattern during training to weight and normalize the count, and outputting the candidate password dictionary according to the probability priority queue during generation, which is suitable for efficient password guessing in offline password audit, password policy evaluation and penetration testing.

The application discloses a VR signal direct connection system and method based on multi-protocol adaptive matching, wherein the VR signal direct connection system comprises a signal receiving unit used for receiving signals of different frequency bands; a protocol decoding unit used for performing protocol decoding on the transmitted signals; an LSTM protocol prediction model used for generating a protocol priority queue and transmitting the generated protocol priority queue to a dynamic protocol loading unit; the dynamic protocol loading unit is used for judging whether to switch protocols according to signal quality conditions and the generated protocol priority queue, and if yes, transmitting the protocol state after switching to an FPGA dynamic partition; the FPGA dynamic partition is used for receiving the protocol state transmitted by the dynamic protocol loading unit and the protocol data transmitted by the protocol decoding unit; a signal processing module is used for receiving the protocol state and the protocol data transmitted by the FPGA dynamic partition, processing the received data information, and generating the content displayed on a VR device.

The application belongs to the technical field of video stream real-time analysis, and particularly relates to a high-concurrency video stream real-time analysis engine scheduling system and method, which comprises a scheduling management layer, a task execution layer and a resource management layer; the scheduling management layer comprises a multi-dimensional task evaluation module, an adaptive scheduling decision module, a multi-level priority queue management module, a scene-aware scheduling module and a dynamic load balancing module; the task execution layer comprises a video stream receiving module, a task decomposition module, a parallel execution engine and a result aggregation module; and the resource management layer comprises a resource monitoring module, a resource allocation module and a resource recycling module. Through the multi-dimensional task evaluation mechanism, the adaptive resource scheduling strategy, the multi-level priority queue, the scene-aware scheduling and the dynamic load balancing, the application realizes low resource consumption, high-concurrency processing and real-time response of video stream analysis scheduling, and significantly improves the video stream real-time analysis performance and resource utilization efficiency in the multi-camera, multi-task and multi-scene conditions.

The invention discloses a multi-point breadth-first search path planning method based on bending radius constraint. The method comprises the following steps: constructing a directed weighted graph model of an environment map; multi-point breadth-first search is carried out, a priority queue containing a plurality of candidate starting points is initialized, and on the basis of the directed weighted graph, the path accumulation bending cost and the edge weight sum serve as search bases, and search is carried out from all the starting points in parallel till a termination condition is met; and backtracking from the target node meeting the termination condition, generating an initial path node sequence, carrying out smooth optimization on the sequence, and outputting a final path meeting the bending radius constraint. According to the method, the dynamic weighted map model is constructed, multi-starting-point parallel breadth-first search is adopted, the bending cost is accumulated in real time, and finally the smooth path conforming to the bending radius constraint is generated through smooth optimization, so that the vehicle driving problem caused by improper path curvature is systematically solved.