Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986842
Hanxiang Gao, Boyang Ren, Quanbo Ge
The In-flight simulator can simulate the flight of the new aircraft before it takes off, and it plays a very important role in improving the design of the new aircraft, and it is an important link in the process of aircraft development. This paper analyzes the cross-linking relationship between the control instructions of the original aircraft and the variable stability aircraft, designs the control mode switching circuit, and gives the overall composition and implementation method of the control mode switching equipment. The gradual loop and double redundancy design are adopted in the design, which effectively ensure the stability and reliability of the equipment operation and realize the switching of the control command of the rudder surface between the original aircraft flight control system and the variable stability flight control system. The test results show that the equipment works well and meets the requirements of flight test.
{"title":"The Design and Implementation of Control Mode Switching Equipment for a Type of In-flight Simulator","authors":"Hanxiang Gao, Boyang Ren, Quanbo Ge","doi":"10.1109/ICUS55513.2022.9986842","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986842","url":null,"abstract":"The In-flight simulator can simulate the flight of the new aircraft before it takes off, and it plays a very important role in improving the design of the new aircraft, and it is an important link in the process of aircraft development. This paper analyzes the cross-linking relationship between the control instructions of the original aircraft and the variable stability aircraft, designs the control mode switching circuit, and gives the overall composition and implementation method of the control mode switching equipment. The gradual loop and double redundancy design are adopted in the design, which effectively ensure the stability and reliability of the equipment operation and realize the switching of the control command of the rudder surface between the original aircraft flight control system and the variable stability flight control system. The test results show that the equipment works well and meets the requirements of flight test.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115307166","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986996
Haonan Zheng, Jiajia Zhao, Wei Tan
In the field of artificial intelligence, neural network is one of the key technologies used for image classification and recognition. However, recent work has demonstrated that deep neural networks are easily attacked by adversarial examples to make misjudgments. Adversarial examples are almost indistin-guishable from normal examples and yet cannot be classified correctly by neural networks. The existence of adversarial examples is a major obstacle to the practical application and deployment of neural networks, so the research on adversarial defense algorithms is an important topic in the field of AI security. This paper proposes an adversarial example defense algorithm based on a triplet network and voting decision mechanism. Firstly, two neural networks with different structures are trained based on normal dataset. Secondly, the first network is fine-tuned using the adversarial examples generated by these two networks, resulting in a third neural network. Then, these three neural networks are used as sub-networks in parallel to construct a triplet network. Through adversarial training and differences in structures, the transferability of adversarial examples among the three sub-networks is weakened. Finally, the final classification result is obtained by majority voting, based on the parallel output results of the three sub-networks. Through the complementarity between these three sub-networks, the defense against adversarial examples is realized. The experimental results demonstrate the effectiveness of this algorithm.
{"title":"An Adversarial Defense Algorithm Based on Triplet Network and Voting Decision","authors":"Haonan Zheng, Jiajia Zhao, Wei Tan","doi":"10.1109/ICUS55513.2022.9986996","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986996","url":null,"abstract":"In the field of artificial intelligence, neural network is one of the key technologies used for image classification and recognition. However, recent work has demonstrated that deep neural networks are easily attacked by adversarial examples to make misjudgments. Adversarial examples are almost indistin-guishable from normal examples and yet cannot be classified correctly by neural networks. The existence of adversarial examples is a major obstacle to the practical application and deployment of neural networks, so the research on adversarial defense algorithms is an important topic in the field of AI security. This paper proposes an adversarial example defense algorithm based on a triplet network and voting decision mechanism. Firstly, two neural networks with different structures are trained based on normal dataset. Secondly, the first network is fine-tuned using the adversarial examples generated by these two networks, resulting in a third neural network. Then, these three neural networks are used as sub-networks in parallel to construct a triplet network. Through adversarial training and differences in structures, the transferability of adversarial examples among the three sub-networks is weakened. Finally, the final classification result is obtained by majority voting, based on the parallel output results of the three sub-networks. Through the complementarity between these three sub-networks, the defense against adversarial examples is realized. The experimental results demonstrate the effectiveness of this algorithm.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116926114","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986602
Q. Tan, Xiaoyang Xie, Tao Yu, Jian Gao
This paper investigates cooperative guidance problems under event-triggered scheme for multiple aircraft with fixed communication topology. The event-triggered distributed cooperative guidance method is divided into two steps. In the first step, an event-triggered scheme is proposed in the cooperative guidance method to reach consensus asymptotically. Each aircraft exchanges the state information with its neighbors at the certain instants when the states satisfy the triggering conditions. The frequency of the communication between the aircraft goes down compared with the periodic communication in the cooperative guidance method. In the second step, each aircraft adopt the widely-used proportional navigation guidance law (PNG) independently to conduct the cooperative mission without communication. Moreover, it is proven that the aircraft can fulfill the cooperative mission asymptotically under the designed triggering condition in the first step. Finally, simulation results are illustrated to verify the validity of the proposed cooperative guidance method under event-triggered scheme.
{"title":"Event-triggered Distributed Cooperative Guidance for Multiple Aircraft","authors":"Q. Tan, Xiaoyang Xie, Tao Yu, Jian Gao","doi":"10.1109/ICUS55513.2022.9986602","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986602","url":null,"abstract":"This paper investigates cooperative guidance problems under event-triggered scheme for multiple aircraft with fixed communication topology. The event-triggered distributed cooperative guidance method is divided into two steps. In the first step, an event-triggered scheme is proposed in the cooperative guidance method to reach consensus asymptotically. Each aircraft exchanges the state information with its neighbors at the certain instants when the states satisfy the triggering conditions. The frequency of the communication between the aircraft goes down compared with the periodic communication in the cooperative guidance method. In the second step, each aircraft adopt the widely-used proportional navigation guidance law (PNG) independently to conduct the cooperative mission without communication. Moreover, it is proven that the aircraft can fulfill the cooperative mission asymptotically under the designed triggering condition in the first step. Finally, simulation results are illustrated to verify the validity of the proposed cooperative guidance method under event-triggered scheme.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121027423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986957
Zhaolei Wang, Kunfeng Lu, Chunmei Yu, Na Yao, Ludi Wang, Jikang Zhao
Aiming at the problem that the traditional control law design process is complex and relies heavily on accurate mathematical models, this paper uses the Deep Deterministic Policy Gradient (DDPG) reinforcement learning to realize the self-learning of continuous motion control law. However, since the performance of the DDPG algorithm depends heavily on the hyper-parameters, there is no clear design basis for the Actor-Critic framework neural network architecture. Considering that the reinforcement learning requires a large amount of computation, the repetitive manual trial and error of hyper-parameters greatly reduces the design efficiency of the algorithm and increases labor costs. On the basis of converting the network architecture design problem into a graph topology generation problem, an automatic search and optimization framework for deep reinforcement learning neural network structure is given in this paper, where the graph topology generation algorithm based on LSTM recurrent neural network, the weight sharing-based lightweight training and evaluation mechanism of deep reinforcement network parameter, and the policy gradient-based learning algorithm of graph topology generator parameter are innovatively combined. Thus, the neural network hyper-parameters in the DDPG algorithm are automatically optimized, and the control law is obtained by self-learning training. Finally, taking rocket vertical recovery control as an ex-ample, the effectiveness of the proposed method is verified.
{"title":"Rocket Self-learning Control based on Lightweight Neural Network Architecture Search","authors":"Zhaolei Wang, Kunfeng Lu, Chunmei Yu, Na Yao, Ludi Wang, Jikang Zhao","doi":"10.1109/ICUS55513.2022.9986957","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986957","url":null,"abstract":"Aiming at the problem that the traditional control law design process is complex and relies heavily on accurate mathematical models, this paper uses the Deep Deterministic Policy Gradient (DDPG) reinforcement learning to realize the self-learning of continuous motion control law. However, since the performance of the DDPG algorithm depends heavily on the hyper-parameters, there is no clear design basis for the Actor-Critic framework neural network architecture. Considering that the reinforcement learning requires a large amount of computation, the repetitive manual trial and error of hyper-parameters greatly reduces the design efficiency of the algorithm and increases labor costs. On the basis of converting the network architecture design problem into a graph topology generation problem, an automatic search and optimization framework for deep reinforcement learning neural network structure is given in this paper, where the graph topology generation algorithm based on LSTM recurrent neural network, the weight sharing-based lightweight training and evaluation mechanism of deep reinforcement network parameter, and the policy gradient-based learning algorithm of graph topology generator parameter are innovatively combined. Thus, the neural network hyper-parameters in the DDPG algorithm are automatically optimized, and the control law is obtained by self-learning training. Finally, taking rocket vertical recovery control as an ex-ample, the effectiveness of the proposed method is verified.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121053481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In order to improve the stability of the dual-rotor Unmanned Aerial Vehicle (UAV), a balance tail is designed to be equipped on the UAV. With reasonable movement, the balance tail may generate additional force and moment which can promote the UAV to be stable rapidly when the UAV is ready to stop. Firstly, the kinematics and dynamics of the tilt-rotor UAV are modeled by Newton-Euler method, and the relations between the movement of the balance tail and the additional force and moment are deduced. The flight control of tilting dual-rotor UAV is realized. Then, the influences of the balance tail on the dual-rotor UAV are analyzed by the nonlinear simulation under the conditions of different masses of the tail and swing rules. Finally, the tail coordinating with the motor tilting and active control of the tail based on cascade PID are explored for stability augmentation of the UAV, respectively. And the effectiveness of the two methods is verified by simulation.
{"title":"Stability Augmentation Control of Tilting Dual-Rotor UAV with Balance Tail","authors":"Yurui Xu, Liang Gao, Benshan Liu, Junming Zhang, Yanhe Zhu, Jie Zhao","doi":"10.1109/ICUS55513.2022.9986615","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986615","url":null,"abstract":"In order to improve the stability of the dual-rotor Unmanned Aerial Vehicle (UAV), a balance tail is designed to be equipped on the UAV. With reasonable movement, the balance tail may generate additional force and moment which can promote the UAV to be stable rapidly when the UAV is ready to stop. Firstly, the kinematics and dynamics of the tilt-rotor UAV are modeled by Newton-Euler method, and the relations between the movement of the balance tail and the additional force and moment are deduced. The flight control of tilting dual-rotor UAV is realized. Then, the influences of the balance tail on the dual-rotor UAV are analyzed by the nonlinear simulation under the conditions of different masses of the tail and swing rules. Finally, the tail coordinating with the motor tilting and active control of the tail based on cascade PID are explored for stability augmentation of the UAV, respectively. And the effectiveness of the two methods is verified by simulation.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124845319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9987057
Jie Xu, Weinan Wu, Yiming Sun
Multi unmanned aerial vehicles (multi-UAVs) coordinated ground attack is an important part of regional sealing and suppression tasks. This paper takes this task as the background to carry out research on multi-UAVs cooperative trajectory planning. Based on this, a multi-UAVs trajectory planning method considering cooperative attack angle and time constraints is proposed. First, the problem is described as a directed graph based on graph theory, the UAV is equivalent to the Dubins Car model, and the constraint model of trajectory avoidance, obstacle avoidance, attack time, and attack angle is given. The total track length is taken as the optimization objective, and the genetic algorithm is designed to solve the problem. The simulation results show that the genetic algorithm can solve the problem, and the designed method has good engineering application value.
{"title":"Multi-UAVs Trajectory Planning Method with Coordinated Attack Angle-Time Constraints","authors":"Jie Xu, Weinan Wu, Yiming Sun","doi":"10.1109/ICUS55513.2022.9987057","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9987057","url":null,"abstract":"Multi unmanned aerial vehicles (multi-UAVs) coordinated ground attack is an important part of regional sealing and suppression tasks. This paper takes this task as the background to carry out research on multi-UAVs cooperative trajectory planning. Based on this, a multi-UAVs trajectory planning method considering cooperative attack angle and time constraints is proposed. First, the problem is described as a directed graph based on graph theory, the UAV is equivalent to the Dubins Car model, and the constraint model of trajectory avoidance, obstacle avoidance, attack time, and attack angle is given. The total track length is taken as the optimization objective, and the genetic algorithm is designed to solve the problem. The simulation results show that the genetic algorithm can solve the problem, and the designed method has good engineering application value.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125147551","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Compared with a short-term task, long-term tracking has received more attention and research in recent years. Long-term tracking is more challenging because it needs to solve two difficult problems: when to update and how to update our model. Many outstanding short-term tracking methods update frame by frame or manually set the threshold to judge if the tracker should be updated, but when the target is blocked or escapes from the field of view, it is easy to get and update wrong samples, resulting in model pollution and drift. Not only that, but due to the lack of a re-detection mechanism, it is difficult for these short-term tracking methods to recover once the target is lost (especially when the target reappears from another location). In this work, we propose a high-speed long-term tracker with adaptive occlusion and recovery judgment (LT-AOR), which comprehensively judge the update chance of the tracker through the discrimination information and appearance information, and re-detects the target in a simplified way to achieve stable tracking in the case of target occlusion and loss.
{"title":"Long-term Tracker with Adaptive Occlusion and Recovery Judgment","authors":"Ying Mi, Chan Liu, Chaohui Wang, Xiangyang Yue, Xiaohan Zhao, Lu Chen","doi":"10.1109/ICUS55513.2022.9986584","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986584","url":null,"abstract":"Compared with a short-term task, long-term tracking has received more attention and research in recent years. Long-term tracking is more challenging because it needs to solve two difficult problems: when to update and how to update our model. Many outstanding short-term tracking methods update frame by frame or manually set the threshold to judge if the tracker should be updated, but when the target is blocked or escapes from the field of view, it is easy to get and update wrong samples, resulting in model pollution and drift. Not only that, but due to the lack of a re-detection mechanism, it is difficult for these short-term tracking methods to recover once the target is lost (especially when the target reappears from another location). In this work, we propose a high-speed long-term tracker with adaptive occlusion and recovery judgment (LT-AOR), which comprehensively judge the update chance of the tracker through the discrimination information and appearance information, and re-detects the target in a simplified way to achieve stable tracking in the case of target occlusion and loss.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116195701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986994
Shuo Pan, Zhiyu Xi
This paper proposes a method for health assessment of unmanned aerial vehicle formation system under false data injection attack. The models of unmanned aerial vehicle and false data injection attack are built, and leader-follower structure is adopted for the formation system. The degree of health reflecting the probability that the error exceeds its allowable value subject to random noise and false data injection attack is established. Derivation of the degree of health is formulated as an optimization problem and procedures to solve the optimization problem are also provided. Finally, simulation is carried out to depict the dynamics of a unmanned aerial vehicle formation system under false data injection attack with the degree of health of the system highlighted.
{"title":"Health Assessment of Unmanned Aerial Vehicle Formation Systems under False Data Injection Attack","authors":"Shuo Pan, Zhiyu Xi","doi":"10.1109/ICUS55513.2022.9986994","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986994","url":null,"abstract":"This paper proposes a method for health assessment of unmanned aerial vehicle formation system under false data injection attack. The models of unmanned aerial vehicle and false data injection attack are built, and leader-follower structure is adopted for the formation system. The degree of health reflecting the probability that the error exceeds its allowable value subject to random noise and false data injection attack is established. Derivation of the degree of health is formulated as an optimization problem and procedures to solve the optimization problem are also provided. Finally, simulation is carried out to depict the dynamics of a unmanned aerial vehicle formation system under false data injection attack with the degree of health of the system highlighted.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"138 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116415825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986811
Liang Zhao, Ruyin Mao, Yong Bai
Unmanned Surface Vehicles (USVs) have witnessed a vigorous growth in the past decades and have been applied in various applications in both commercial and military domains. Central to the control of USVs, local path planning is one of the crucial technologies in the process towards autonomy. This paper investigates the application of a novel path planning algorithm in combination with modified Hybrid A* and several path manipulators. First, the heuristic strategy and node expansion method of standard Hybrid A* have been adapted for the algorithm. Second, node cutting technique and B-spline path smoother are applied to enhance the path quality. Simulations have been carried out to illustrate the excellent performance of the proposed method comparing to several state-of-the-art methods. Furthermore, the new algorithm is tested in the USV model, and the results have demonstrated that it can perfectly coordinate with the USV control system. Therefore, the proposed algorithm can be considered as a reliable method dealing with local path planning problem for USV.
无人水面飞行器(usv)在过去的几十年里得到了蓬勃的发展，并在商业和军事领域得到了广泛的应用。在无人驾驶汽车的控制中，局部路径规划是实现自动驾驶的关键技术之一。本文研究了一种结合改进的Hybrid a *和多个路径操纵器的新型路径规划算法的应用。首先，采用了标准Hybrid A*的启发式策略和节点展开方法。其次，采用节点切割技术和b样条路径平滑来提高路径质量;仿真表明，与几种最先进的方法相比，所提出的方法具有优异的性能。此外，在USV模型中对该算法进行了测试，结果表明该算法能够很好地与USV控制系统协调。因此，该算法是解决无人潜航器局部路径规划问题的可靠方法。
{"title":"Local Path Planning for Unmanned Surface Vehicles based on Hybrid A* and B-spline","authors":"Liang Zhao, Ruyin Mao, Yong Bai","doi":"10.1109/ICUS55513.2022.9986811","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986811","url":null,"abstract":"Unmanned Surface Vehicles (USVs) have witnessed a vigorous growth in the past decades and have been applied in various applications in both commercial and military domains. Central to the control of USVs, local path planning is one of the crucial technologies in the process towards autonomy. This paper investigates the application of a novel path planning algorithm in combination with modified Hybrid A* and several path manipulators. First, the heuristic strategy and node expansion method of standard Hybrid A* have been adapted for the algorithm. Second, node cutting technique and B-spline path smoother are applied to enhance the path quality. Simulations have been carried out to illustrate the excellent performance of the proposed method comparing to several state-of-the-art methods. Furthermore, the new algorithm is tested in the USV model, and the results have demonstrated that it can perfectly coordinate with the USV control system. Therefore, the proposed algorithm can be considered as a reliable method dealing with local path planning problem for USV.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116527317","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-10-28DOI: 10.1109/ICUS55513.2022.9986793
Chengbo Wang, Xinyu Zhang, Hongbo Gao, Huiping Su, Kangjie Zheng, Weisong Wang
In this paper, a learning experience reuse - reinforcement learning collision avoidance (LER-RLCA) method is proposed, which can synthesize near-optimal collision avoidance policy with efficient sampling and good seamanship, to solve the local safety sailing of autonomous ship in a multi-obstacle environment. Lying on the general reinforcement learning (RL), using learning experience reuse, the hidden features of historical training data were mined. Meanwhile, a new reward function combining external revenue signal with internal incentive signal was designed to encourage search the environment with a low probability of state transition. We further applied LER-RLCA algorithm to the simulation of autonomous ship collision avoidance. The results show that the proposed LER-RLCA algorithm can well realize the collision-free and safe navigation of autonomous ships, to avoid falling into local iteration, greatly improve the convergence speed of the algorithm, and improve the performance of online collision avoidance decision-making.
{"title":"Efficient Reinforcement Learning for Autonomous Ship Collision Avoidance under Learning Experience Reuse","authors":"Chengbo Wang, Xinyu Zhang, Hongbo Gao, Huiping Su, Kangjie Zheng, Weisong Wang","doi":"10.1109/ICUS55513.2022.9986793","DOIUrl":"https://doi.org/10.1109/ICUS55513.2022.9986793","url":null,"abstract":"In this paper, a learning experience reuse - reinforcement learning collision avoidance (LER-RLCA) method is proposed, which can synthesize near-optimal collision avoidance policy with efficient sampling and good seamanship, to solve the local safety sailing of autonomous ship in a multi-obstacle environment. Lying on the general reinforcement learning (RL), using learning experience reuse, the hidden features of historical training data were mined. Meanwhile, a new reward function combining external revenue signal with internal incentive signal was designed to encourage search the environment with a low probability of state transition. We further applied LER-RLCA algorithm to the simulation of autonomous ship collision avoidance. The results show that the proposed LER-RLCA algorithm can well realize the collision-free and safe navigation of autonomous ships, to avoid falling into local iteration, greatly improve the convergence speed of the algorithm, and improve the performance of online collision avoidance decision-making.","PeriodicalId":345773,"journal":{"name":"2022 IEEE International Conference on Unmanned Systems (ICUS)","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116534931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：