Real-time UAV path planning based on LSTM network

To address the shortcomings of single-step decision making in the existing deep reinforcement learning based unmanned aerial vehicle(UAV)real-time path planning problem,a real-time UAV path planning algorithm based on long shortterm memory(RPP-LSTM)network is proposed,which combines the memory characteristics of recurrent neural network(RNN)and the deep reinforcement learning algorithm.LSTM networks are used in this algorithm as Q-value networks for the deep Q network(DQN)algorithm,which makes the decision of the Q-value network has some memory.Thanks to LSTM network,the Q-value network can use the previous environmental information and action information which effectively avoids the problem of single-step decision considering only the current environment.Besides,the algorithm proposes a hierarchical reward and punishment function for the specific problem of UAV real-time path planning,so that the UAV can more reasonably perform path planning.Simulation verification shows that compared with the traditional feed-forward neural network(FNN)based UAV autonomous path planning algorithm,the RPP-LSTM proposed in this paper can adapt to more complex environments and has significantly improved robustness and accuracy when performing UAV real-time path planning.

Improved Double Deep Q Network Algorithm Based on Average Q-Value Estimation and Reward Redistribution for Robot Path Planning

By integrating deep neural networks with reinforcement learning,the Double Deep Q Network(DDQN)algorithm overcomes the limitations of Q-learning in handling continuous spaces and is widely applied in the path planning of mobile robots.However,the traditional DDQN algorithm suffers from sparse rewards and inefficient utilization of high-quality data.Targeting those problems,an improved DDQN algorithm based on average Q-value estimation and reward redistribution was proposed.First,to enhance the precision of the target Q-value,the average of multiple previously learned Q-values from the target Q network is used to replace the single Q-value from the current target Q network.Next,a reward redistribution mechanism is designed to overcome the sparse reward problem by adjusting the final reward of each action using the round reward from trajectory information.Additionally,a reward-prioritized experience selection method is introduced,which ranks experience samples according to reward values to ensure frequent utilization of high-quality data.Finally,simulation experiments are conducted to verify the effectiveness of the proposed algorithm in fixed-position scenario and random environments.The experimental results show that compared to the traditional DDQN algorithm,the proposed algorithm achieves shorter average running time,higher average return and fewer average steps.The performance of the proposed algorithm is improved by 11.43%in the fixed scenario and 8.33%in random environments.It not only plans economic and safe paths but also significantly improves efficiency and generalization in path planning,making it suitable for widespread application in autonomous navigation and industrial automation.

Motion Planning for Autonomous Driving with Real Traffic Data Validation

Accurate trajectory prediction of surrounding road users is the fundamental input for motion planning,which enables safe autonomous driving on public roads.In this paper,a safe motion planning approach is proposed based on the deep learning-based trajectory prediction method.To begin with,a trajectory prediction model is established based on the graph neural network(GNN)that is trained utilizing the INTERACTION dataset.Then,the validated trajectory prediction model is used to predict the future trajectories of surrounding road users,including pedestrians and vehicles.In addition,a GNN prediction model-enabled motion planner is developed based on the model predictive control technique.Furthermore,two driving scenarios are extracted from the INTERACTION dataset to validate and evaluate the effectiveness of the proposed motion planning approach,i.e.,merging and roundabout scenarios.The results demonstrate that the proposed method can lower the risk and improve driving safety compared with the baseline method.

Planning,monitoring and replanning techniques for handling abnormity in HTN-based planning and execution

A framework that integrates planning,monitoring and replanning techniques is proposed.It can devise the best solution based on the current state according to specific objectives and properly deal with the influence of abnormity on the plan execution.The framework consists of three parts:the hierarchical task network(HTN)planner based on Monte Carlo tree search(MCTS),hybrid plan monitoring based on forward and backward and norm-based replanning method selection.The HTN planner based on MCTS selects the optimal method for HTN compound task through pre-exploration.Based on specific objectives,it can identify the best solution to the current problem.The hybrid plan monitoring has the capability to detect the influence of abnormity on the effect of an executed action and the premise of an unexecuted action,thus trigger the replanning.The norm-based replanning selection method can measure the difference between the expected state and the actual state,and then select the best replanning algorithm.The experimental results reveal that our method can effectively deal with the influence of abnormity on the implementation of the plan and achieve the target task in an optimal way.

Correlation knowledge extraction based on data mining for distribution network planning

Traditional distribution network planning relies on the professional knowledge of planners,especially when analyzing the correlations between the problems existing in the network and the crucial influencing factors.The inherent laws reflected by the historical data of the distribution network are ignored,which affects the objectivity of the planning scheme.In this study,to improve the efficiency and accuracy of distribution network planning,the characteristics of distribution network data were extracted using a data-mining technique,and correlation knowledge of existing problems in the network was obtained.A data-mining model based on correlation rules was established.The inputs of the model were the electrical characteristic indices screened using the gray correlation method.The Apriori algorithm was used to extract correlation knowledge from the operational data of the distribution network and obtain strong correlation rules.Degree of promotion and chi-square tests were used to verify the rationality of the strong correlation rules of the model output.In this study,the correlation relationship between heavy load or overload problems of distribution network feeders in different regions and related characteristic indices was determined,and the confidence of the correlation rules was obtained.These results can provide an effective basis for the formulation of a distribution network planning scheme.

Coordinated Planning Transmission Tasks in Heterogeneous Space Networks:A Semi-Distributed Approach

This paper studies the coordinated planning of transmission tasks in the heterogeneous space networks to enable efficient sharing of ground stations cross satellite systems.Specifically,we first formulate the coordinated planning problem into a mixed integer liner programming(MILP)problem based on time expanded graph.Then,the problem is transferred and reformulated into a consensus optimization framework which can be solved by satellite systems parallelly.With alternating direction method of multipliers(ADMM),a semi-distributed coordinated transmission task planning algorithm is proposed,in which each satellite system plans its own tasks based on local information and limited communication with the coordination center.Simulation results demonstrate that compared with the centralized and fully-distributed methods,the proposed semi-distributed coordinated method can strike a better balance among task complete rate,complexity,and the amount of information required to be exchanged.

Information Freshness-Oriented Trajectory Planning and Resource Allocation for UAV-Assisted Vehicular Networks

In this paper,multi-UAV trajectory planning and resource allocation are jointly investigated to improve the information freshness for vehicular networks,where the vehicles collect time-critical traffic information by on-board sensors and upload to the UAVs through their allocated spectrum resource.We adopt the expected sum age of information(ESAoI)to measure the network-wide information freshness.ESAoI is jointly affected by both the UAVs trajectory and the resource allocation,which are coupled with each other and make the analysis of ESAoI challenging.To tackle this challenge,we introduce a joint trajectory planning and resource allocation procedure,where the UAVs firstly fly to their destinations and then hover to allocate resource blocks(RBs)during a time-slot.Based on this procedure,we formulate a trajectory planning and resource allocation problem for ESAoI minimization.To solve the mixed integer nonlinear programming(MINLP)problem with hybrid decision variables,we propose a TD3 trajectory planning and Round-robin resource allocation(TTPRRA).Specifically,we exploit the exploration and learning ability of the twin delayed deep deterministic policy gradient algorithm(TD3)for UAVs trajectory planning,and utilize Round Robin rule for the optimal resource allocation.With TTP-RRA,the UAVs obtain their flight velocities by sensing the locations and the age of information(AoI)of the vehicles,then allocate the RBs to the vehicles in a descending order of AoI until the remaining RBs are not sufficient to support another successful uploading.Simulation results demonstrate that TTP-RRA outperforms the baseline approaches in terms of ESAoI and average AoI(AAoI).

An Integrated Image Task Planning in Satellite Networks:From Instruction Release and Observation Perspective

The unreasonable observation arrangements in the satellite operation control center(SOCC)may result in the observation data cannot be downloaded as scheduled.Meanwhile,if the operation instructions released by the satellite telemetry tracking center(STTC)for the on-board payloads are not injected on the specific satellites in time,the corresponding satellites cannot perform the observation operations as planned.Therefore,there is an urgent need to design an integrated instruction release,and observation task planning(I-IRO-TP)scheme by efficiently collaborating the SOCC and STTC.Motivated by this fact,we design an interaction mechanism between the SOCC and the STTC,where we first formulate the I-IRO-TP problem as a constraint satisfaction problem aiming at maximizing the number of completed tasks.Furthermore,we propose an interactive imaging task planning algorithm based on the analysis of resource distribution in the STTC during the previous planning periods to preferentially select the observation arcs that not only satisfy the requirements in the observation resource allocation phase but also facilitate the arrangement of measurement and control instruction release.We conduct extensive simulations to demonstrate the effectiveness of the proposed algorithm in terms of the number of completed tasks.

Technical Research on Characteristics and Network Planning of Railway 5G-R Network

The railway 5G-R network is ready for the engineering construction.In order to provide technical reserve for the preliminary engineering stage,it is necessary to study the characteristics of 5G-R network and network planning technology in advance.The 5G-R network,different from the operators'5G network,is mainly available for railway staff and covers the railway line.It is upgraded based on the previous generation of railway communication system-GSM-R technology.By comprehensively summarizing and reviewing the characteristics of 5G-R system architecture,network coverage,networking technology,business characteristics,equipment redundancy,network performance,etc.,we study the network planning technology with 5G-R characteristics based on the above characteristics,including frequency,coverage,capacity,parameters,indoor distribution,switching and other key technologies.On this basis,this paper puts forward 5G-R network planning suggestion,which can provide technical support for subsequent project deployment.

Research on Grid Planning of Dual Power Distribution Network Based on Parallel Ant Colony Optimization Algorithm

A distribution network plays an extremely important role in the safe and efficient operation of a power grid.As the core part of a power grid’s operation,a distribution network will have a significant impact on the safety and reliability of residential electricity consumption.it is necessary to actively plan and modify the distribution network’s structure in the power grid,improve the quality of the distribution network,and optimize the planning of the distribution network,so that the network can be fully utilized to meet the needs of electricity consumption.In this paper,a distribution network grid planning algorithm based on the reliability of electricity consumption was completed using ant colony algorithm.For the distribution network structure planning of dual power sources,the parallel ant colony algorithm was used to prove that the premise of parallelism is the interactive process of ant colonies,and the dual power distribution network structure model is established based on the principle of the lowest cost.The artificial ants in the algorithm were compared with real ants in nature,and the basic steps and working principle of the ant colony optimization algorithm was studied with the help of the travelling salesman problem(TSP).Then,the limitations of the ant colony algorithm were analyzed,and an improvement strategy was proposed by using python for digital simulation.The results demonstrated the reliability of model-building and algorithm improvement.

Design of Aided Decision-Making Program for Prioritizing Construction Projects in Urban Road Network Planning

The importance and complexity of prioritizing construction projects (PCP) in urban road network planning lead to the necessity to develop an aided decision making program (ADMP). Cost benefit ratio model and stage rolled method are chosen as the theoretical foundations of the program, and then benefit model is improved to accord with the actuality of urban traffic in China. Consequently, program flows, module functions and data structures are designed, and particularly an original data structure of road ...

Specific Problems in the Planning and Development of Power Network in China

In light of the situation that the nationwide interconnection of power networks in China in the coming years will take shape, it is imperative to emphasize the importance of setting up rational power network configuration. Combined with the characteristics of regional power networks in China, problems in network planning that need to be solved are put forward in this paper, such as, the access of power plants to grid by layers and zones, the share of external power in the load of local network, the power network configuration study in-depth in planning and design stage, and enforcement of receiving-end power network trunk etc. The background of these problems and their countermeasures are also analyzed in the paper.

Reuse partitioning based frequency planning for two-hop cellular network

Following the principle of reuse partitioning, two new frequency planning schemes are proposed, the coverage-oriented scheme and the efficiency-oriented scheme, for the cellular system with two-hop fixed relay nodes （FRNs）. Compared with the effficiency-oriented scheme, the coverage-oriented scheme has higher reuse distances and is developed with emphasis on the coverage, while compared with the coverage-oriented scheme, the efficiency-oriented scheme has smaller reuse distances and is developed with emphasis on the spectral efficiency. Taking uplink as an example, both simplified analysis and intensive computer simulations are presented to offer comparisons among FRN enhanced systems with the proposed schemes, with a known channel-borrowing based frequency planning scheme and the conventional cellular system without relaying. Studies show that the FRN enhanced system with the coverage-oriented scheme provides the best coverage, while that with the efficiency-oriented scheme offers the highest area spectral efficiency.

Generative Adversarial Network Based Heuristics for Sampling-Based Path Planning

Sampling-based path planning is a popular methodology for robot path planning.With a uniform sampling strategy to explore the state space,a feasible path can be found without the complex geometric modeling of the configuration space.However,the quality of the initial solution is not guaranteed,and the convergence speed to the optimal solution is slow.In this paper,we present a novel image-based path planning algorithm to overcome these limitations.Specifically,a generative adversarial network(GAN)is designed to take the environment map(denoted as RGB image)as the input without other preprocessing works.The output is also an RGB image where the promising region(where a feasible path probably exists)is segmented.This promising region is utilized as a heuristic to achieve non-uniform sampling for the path planner.We conduct a number of simulation experiments to validate the effectiveness of the proposed method,and the results demonstrate that our method performs much better in terms of the quality of the initial solution and the convergence speed to the optimal solution.Furthermore,apart from the environments similar to the training set,our method also works well on the environments which are very different from the training set.

A hybrid particle swarm optimization approach with neural network and set pair analysis for transmission network planning

Transmission network planning (TNP) is a large-scale, complex, with more non-linear discrete variables and the multi-objective constrained optimization problem. In the optimization process, the line investment, network reliability and the network loss are the main objective of transmission network planning. Combined with set pair analysis (SPA), particle swarm optimization (PSO), neural network (NN), a hybrid particle swarm optimization model was established with neural network and set pair analysis for transmission network planning (HPNS). Firstly, the contact degree of set pair analysis was introduced, the traditional goal set was converted into the collection of the three indicators including the identity degree, difference agree and contrary degree. On this bases, using shi(H), the three objective optimization problem was converted into single objective optimization problem. Secondly, using the fast and efficient search capabilities of PSO, the transmission network planning model based on set pair analysis was optimized. In the process of optimization, by improving the BP neural network constantly training so that the value of the fitness function of PSO becomes smaller in order to obtain the optimization program fitting the three objectives better. Finally, compared HPNS with PSO algorithm and the classic genetic algorithm, HPNS increased about 23% efficiency than THA, raised about 3.7% than PSO and improved about 2.96% than GA.

Path Planning and Tracking for Vehicle Parallel Parking Based on Preview BP Neural Network PID Controller

In order to diminish the impacts of extemal disturbance such as parking speed fluctuation and model un- certainty existing in steering kinematics, this paper presents a parallel path tracking method for vehicle based on pre- view back propagation （BP） neural network PID controller. The forward BP neural network can adjust the parameters of PID controller in real time. The preview time is optimized by considering path curvature, change in curvature and road boundaries. A fuzzy controller considering barriers and different road conditions is built to select the starting po- sition. In addition, a kind of path planning technology satisfying the requirement of obstacle avoidance is introduced. In order to solve the problem of discontinuous curvature, cubic B spline curve is used for curve fitting. The simulation results and real vehicle tests validate the effectiveness of the proposed path planning and tracking methods.

A Novel MILP Model Based on the Topology of a Network Graph for Process Planning in an Intelligent Manufacturing System

Intelligent process planning(PP)is one of the most important components in an intelligent manufacturing system and acts as a bridge between product designing and practical manufacturing.PP is a nondeterministic polynomial-time(NP)-hard problem and,as existing mathematical models are not formulated in linear forms,they cannot be solved well to achieve exact solutions for PP problems.This paper proposes a novel mixed-integer linear programming(MILP)mathematical model by considering the network topology structure and the OR nodes that represent a type of OR logic inside the network.Precedence relationships between operations are discussed by raising three types of precedence relationship matrices.Furthermore,the proposed model can be programmed in commonly-used mathematical programming solvers,such as CPLEX,Gurobi,and so forth,to search for optimal solutions for most open problems.To verify the effectiveness and generality of the proposed model,five groups of numerical experiments are conducted on well-known benchmarks.The results show that the proposed model can solve PP problems effectively and can obtain better solutions than those obtained by the state-ofthe-art algorithms.

Analysis of effect factors-based stochastic network planning model

Looking at all the indeterminate factors as a whole and regarding activity durations as independent random variables, the traditional stochastic network planning models ignore the inevitable relationship and dependence among activity durations when more than one activity is possibly affected by the same indeterminate factors. On this basis of analysis of indeterminate effect factors of durations, the effect factors-based stochastic network planning （EFBSNP） model is proposed, which emphasizes on the effects of not only logistic and organizational relationships, but also the dependent relationships, due to indeterminate factors among activity durations on the project period. By virtue of indeterminate factor analysis the model extracts and describes the quantitatively indeterminate effect factors, and then takes into account the indeterminate factors effect schedule by using the Monte Carlo simulation technique. The method is flexible enough to deal with effect factors and is coincident with practice. A software has been developed to simplify the model-based calculation, in VisualStudio.NET language. Finally, a case study is included to demonstrate the applicability of the proposed model and comparison is made with some advantages over the existing models.

Application and Realization of Hierarchical Network Planning Based on Web

Making use of Microsoft Visual Studio. NET platform, hierarchical network planning is realized in working procedure time-optimization of the construction by TBM, and hierarchical network graph of the construction by TBM is drawn based on browser. Then the theory of system realization is discussed, six components of system that can be reused are explained emphatically. The realization of hierarchical network panning in Internet provides available guarantee for controlling rate of progress in large-scale or middle-sized projects. Key words Web - network graph - hierarchical network planning CLC number TU721 Biography: WU Shi-jing (1963-), male, Professor, research direction: equipments management engineering, mechatronics, state monitoring and malfunction diagnostics of machine and electronics equipments.

Integration of Neural Networks and Cellular Automata for Urban Planning

This paper presents a new type of cellular automa ta (CA) model for the simulation of alternative land development using neural netw orks for urban planning. CA models can be regarded as a planning tool because th ey can generate alternative urban growth. Alternative development patterns can b e formed by using different sets of parameter values in CA simulation. A critica l issue is how to define parameter values for realistic and idealized simulation . This paper demonstrates that neural networks can simplify CA models but genera te more plausible results. The simulation is based on a simple three-layer netw ork with an output neuron to generate conversion probability. No transition rule s are required for the simulation. Parameter values are automatically obtained f rom the training of network by using satellite remote sensing data. Original tra ining data can be assessed and modified according to planning objectives. Altern ative urban patterns can be easily formulated by using the modified training dat a sets rather than changing the model.