Trajectory planning method is a research hotspot in autonomous driving. Existing reinforcement learning-based trajectory planning methods suffer from unstable performance due to the strong randomness of network weight...
详细信息
Trajectory planning method is a research hotspot in autonomous driving. Existing reinforcement learning-based trajectory planning methods suffer from unstable performance due to the strong randomness of network weight parameter updates during the training process. Therefore, this paper proposes a novel trajectory planning method based on deep reinforcement learning trust region policy optimization (TRPO). Firstly, in order to enhance the robustness of the trajectory planning method based on deep reinforcement learning TRPO, a TRPO-LSTM based decision model was proposed. More specifically, a long short term memory (LSTM) based state feature extraction network was designed and embeded into a TRPO-based decision model to enhance the ability of TRPO to extract information from the environmental state space. Secondly, in order to make the planned trajectory adaptive to the dynamic changes of traffic environment, we presented a novel TRPO-LSTM trajectory fitting algorithm. To the best of our knowledge, this is the first work aiming at applying the TRPO-LSTM based decision model in the trajectory fitting process to search the optimal longitudinal trajectory speed. Finally, the proposed trajectory planning method was implemented and simulated on the CARLA simulator. The experimental results show that, compared with existing trajectory planning methods based on deep reinforcement learning algorithms, our proposed method achieves a cumulative reward improvement of over 28.9% in the scenario of four lane highway, and has better robustness. Meanwhile, the proposed method can achieve a lower collision rate of 0.93% while improving the average speed and comfort of vehicle driving. IEEE
Grasp detection is a visual recognition task where the robot makes use of its sensors to detect graspable objects in its *** the steady progress in robotic grasping,it is still difficult to achieve both real-time and ...
详细信息
Grasp detection is a visual recognition task where the robot makes use of its sensors to detect graspable objects in its *** the steady progress in robotic grasping,it is still difficult to achieve both real-time and high accuracy grasping *** this paper,we propose a real-time robotic grasp detection method,which can accurately predict potential grasp for parallel-plate robotic grippers using RGB *** work employs an end-to-end convolutional neural network which consists of a feature descriptor and a grasp *** for the first time,we add an attention mechanism to the grasp detection task,which enables the network to focus on grasp regions rather than ***,we present an angular label smoothing strategy in our grasp detection method to enhance the fault tolerance of the *** quantitatively and qualitatively evaluate our grasp detection method from different aspects on the public Cornell dataset and Jacquard *** experiments demonstrate that our grasp detection method achieves superior performance to the state-of-the-art *** particular,our grasp detection method ranked first on both the Cornell dataset and the Jacquard dataset,giving rise to the accuracy of 98.9%and 95.6%,respectively at realtime calculation speed.
Aiming at the problem that bounding boxes need to be defined manually in Unity real-time interactive program, an automatic generation algorithm of hierarchical bounding boxes is proposed. Firstly, the advantages and f...
详细信息
Due to the complex flow state of pneumatically conveyed particles and the influence of the conveying conditions, existing measurement techniques have limitations in detecting the dynamic parameters of full-sections pa...
详细信息
In order to address the issue of low segmentation accuracy in the weak flame region of waste incineration flame images and the potential loss of texture details at the flame edge, this study proposes an algorithm for ...
详细信息
Querying power big data across multiple sections plays an important role in achieving efficient power data sharing and emergent accident handling. However, power big data often suffers from the issues of data sparsity...
详细信息
In order to improve the mean average precision (mAP) of scene image retrieval, this paper proposes a scene image retrieval algorithm based on salient local feature aggregation and geographic information, and establish...
详细信息
For multiagent systems with switching topology, the data-driven fault-tolerant consensus is investigated in this work. The actuator faults are directly estimated using the proposed enhanced-RBFNN-based fault estimatio...
详细信息
This paper is concerned with the multi-agent systems with both packet dropout and input delay.A novel receding horizon control(RHC)based consensus protocol is proposed by solving a distributed RHC based optimization *...
详细信息
This paper is concerned with the multi-agent systems with both packet dropout and input delay.A novel receding horizon control(RHC)based consensus protocol is proposed by solving a distributed RHC based optimization *** novelty of the optimization problem lines in the involvement of the neighbours’predictor information in the cost *** on the derived RHC based consensus protocol,the necessary and sufficient condition for the mean-square consensus is *** addition,the authors give a specific sufficient condition to guarantee the mean-square consensus.
Graph neural network (GNN) has gained increasing popularity in recent years owing to its capability and flexibility in modeling complex graph structure data. Among all graph learning methods, hypergraph learning is a ...
详细信息
暂无评论