9 December 2022 in IEEE Transactions on Instrumentation and Measurement
Simultaneous Localization and Mapping (SLAM) is one of the fundamental capabilities for intelligent mobile robots to perform state estimation in unknown environments. However, most visual SLAM systems rely on the static scene assumption and consequently have severely reduced accuracy and robustness in dynamic scenes. Moreover, the metric maps constructed by many systems lack semantic information, so the robots cannot understand their surroundings at a human cognitive level. In this paper, we propose SG-SLAM, which is a real-time RGB-D semantic visual SLAM system based on the ORB-SLAM2 framework. First, SG-SLAM adds two new parallel threads: an object detecting thread to obtain 2D semantic information and a semantic mapping thread. Then, a fast dynamic feature rejection algorithm combining semantic and geometric information is added to the tracking thread. Finally, they are published to the ROS system for visualization after generating 3D point clouds and 3D semantic objects in the semantic mapping thread. We performed an experimental evaluation on the TUM dataset, the Bonn dataset, and the OpenLORIS-Scene dataset. The results show that SG-SLAM is not only one of the most real-time, accurate, and robust systems in dynamic scenes, but also allows the creation of intuitive semantic metric maps.
28 September 2013 in Chinese Journal of Mechanical Engineering
It is desired to require a walking robot for the elderly and the disabled to have large capacity, high stiffness, stability, etc. However, the existing walking robots cannot achieve these requirements because of the weight-payload ratio and simple function. Therefore, Improvement of enhancing capacity and functions of the walking robot is an important research issue. According to walking requirements and combining modularization and reconfigurable ideas, a quadruped/biped reconfigurable walking robot with parallel leg mechanism is proposed. The proposed robot can be used for both a biped and a quadruped walking robot. The kinematics and performance analysis of a 3-UPU parallel mechanism which is the basic leg mechanism of a quadruped walking robot are conducted and the structural parameters are optimized. The results show that performance of the walking robot is optimal when the circumradius R, r of the upper and lower platform of leg mechanism are 161.7 mm, 57.7 mm, respectively. Based on the optimal results, the kinematics and dynamics of the quadruped walking robot in the static walking mode are derived with the application of parallel mechanism and influence coefficient theory, and the optimal coordination distribution of the dynamic load for the quadruped walking robot with over-determinate inputs is analyzed, which solves dynamic load coupling caused by the branches’ constraint of the robot in the walk process. Besides laying a theoretical foundation for development of the prototype, the kinematics and dynamics studies on the quadruped walking robot also boost the theoretical research of the quadruped walking and the practical applications of parallel mechanism.
20 January 2023 in IEEE Sensors Journal
Human waist action recognition based on surface EMG (Electromyography) signals is a complex pattern recognition problem, the difficulty of which lies in the coupling of EMG signals between different actions, and a single model cannot effectively utilize the feature information to achieve high accuracy of action recognition. To this end, this paper proposes a Stacking integration-based EMG signal model for the waist, which introduces the Stacking integration framework into EMG signal recognition for the first time, using KNN (K-Nearest Neighbor), RF (Random Forest-master), XGboost, and LightGBM algorithms as base learners and DT (Decision tree) model as a meta-learner to construct the integrated model. The complex feature information contained in the EMG signal is correlated with different waist movements to construct feature datasets with high correlation, and the training set is divided by using the five-fold cross-validation method to reduce the overfitting problem during the repeated training of the integrated model. Using the integrated model proposed in this paper, the generalization ability of the model can be effectively improved and the recognition accuracy of EMG signals can be improved. According to experiments, the model has an average accuracy of 94.98% for the recognition of six waist movements in normal person, and the accuracy of recognition of the six movements in patients with lumbar strain was 91.5% on average, which is about 6% better than a single model, and the recognition speed can be within 150ms, meeting the requirements of real-time recognition.
1 October 2013 in International Journal of Automation and Computing