Gang Peng

Nikita Gordienko, Peng Gang, Yuri Gordienko et al.

Machine learning techniques are presented for automatic recognition of the historical letters (XI-XVIII centuries) carved on the stoned walls of St.Sophia cathedral in Kyiv (Ukraine). A new image dataset of these carved Glagolitic and Cyrillic letters (CGCL) was assembled and pre-processed for recognition and prediction by machine learning methods. The dataset consists of more than 4000 images for 34 types of letters. The explanatory data analysis of CGCL and notMNIST datasets shown that the carved letters can hardly be differentiated by dimensionality reduction methods, for example, by t-distributed stochastic neighbor embedding (tSNE) due to the worse letter representation by stone carving in comparison to hand writing. The multinomial logistic regression (MLR) and a 2D convolutional neural network (CNN) models were applied. The MLR model demonstrated the area under curve (AUC) values for receiver operating characteristic (ROC) are not lower than 0.92 and 0.60 for notMNIST and CGCL, respectively. The CNN model gave AUC values close to 0.99 for both notMNIST and CGCL (despite the much smaller size and quality of CGCL in comparison to notMNIST) under condition of the high lossy data augmentation. CGCL dataset was published to be available for the data science community as an open source resource.

Gang Peng, Jin Yang, Xinde Lia et al.

(This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.) To improve the efficiency of deep reinforcement learning (DRL)-based methods for robot manipulator trajectory planning in random working environments, we present three dense reward functions. These rewards differ from the traditional sparse reward. First, a posture reward function is proposed to speed up the learning process with a more reasonable trajectory by modeling the distance and direction constraints, which can reduce the blindness of exploration. Second, a stride reward function is proposed to improve the stability of the learning process by modeling the distance and movement distance of joint constraints. Finally, in order to further improve learning efficiency, we are inspired by the cognitive process of human behavior and propose a stage incentive mechanism, including a hard stage incentive reward function and a soft stage incentive reward function. Extensive experiments show that the soft stage incentive reward function is able to improve the convergence rate by up to 46.9% with the state-of-the-art DRL methods. The percentage increase in the convergence mean reward was 4.4-15.5% and the percentage decreases with respect to standard deviation were 21.9-63.2%. In the evaluation experiments, the success rate of trajectory planning for a robot manipulator reached 99.6%.

Peng Gang, Wang Zhihao, Yang Jin et al.

The traditional kinematic calibration method for manipulators requires precise three-dimensional measuring instruments to measure the end pose, which is not only expensive due to the high cost of the measuring instruments but also not applicable to all manipulators. Another calibration method uses a camera, but the system error caused by the camera's parameters affects the calibration accuracy of the kinematics of the robot arm. Therefore, this paper proposes a method for calibrating the geometric parameters of a kinematic model of a manipulator based on monocular vision. Firstly, the classic Denavit-Hartenberg(D-H) modeling method is used to establish the kinematic parameters of the manipulator. Secondly, nonlinear optimization and parameter compensation are performed. The three-dimensional positions of the feature points of the calibration plate under each manipulator attitude corresponding to the actual kinematic model and the classic D-H kinematic model are mapped into the pixel coordinate system, and the sum of Euclidean distance errors of the pixel coordinates of the two is used as the objective function to be optimized. The experimental results show that the pixel deviation of the end pose corresponding to the optimized D-H kinematic model proposed in this paper and the end pose corresponding to the actual kinematic model in the pixel coordinate system is 0.99 pixels. Compared with the 7.9 deviation pixels between the pixel coordinates calculated by the classic D-H kinematic model and the actual pixel coordinates, the deviation is reduced by nearly 7 pixels for an 87% reduction in error. Therefore, the proposed method can effectively avoid system errors caused by camera parameters in visual calibration, can improve the absolute positioning accuracy of the end of the robotic arm, and has good economy and universality.

Gang Peng, Zezao Lu, Zejie Tan et al.

Mobile robot positioning, mapping, and navigation systems generally employ an inertial measurement unit (IMU) to obtain the acceleration and angular velocity of the robot. However, errors in the internal and external parameters of an IMU arising from defective calibration directly affect the accuracy of robot positioning and pose estimation. While this issue has been addressed by the mature internal reference calibration methods available for IMUs, external reference calibration methods between the IMU and the chassis of a mobile robot are lacking. This study addresses this issue by proposing a novel chassis-IMU internal and external parameter calibration algorithm based on nonlinear optimization, which is designed for robots equipped with cameras, IMUs, and wheel speed odometers, and functions under the premise of accurate calibrations for the internal parameters of the IMU and the internal and external parameters of the camera. All of the internal and external reference calibrations are conducted using the robot's existing equipment without the need for additional calibration aids. The feasibility of the method is verified by its application to a Mecanum wheel omnidirectional mobile platform as an example, as well as suitable for other type chassis of mobile robots. The proposed calibration method is thereby demonstrated to guarantee the accuracy of robot pose estimation.

Gang Peng, Yuezhi Zheng, Jianfeng Li et al.

At present, most high-accuracy single-person pose estimation methods have high computational complexity and insufficient real-time performance due to the complex structure of the network model. However, a single-person pose estimation method with high real-time performance also needs to improve its accuracy due to the simple structure of the network model. It is currently difficult to achieve both high accuracy and real-time performance in single-person pose estimation. For use in human-machine cooperative operations, this paper proposes a single-person upper limb pose estimation method based on an end-to-end approach for accurate and real-time limb pose estimation. Using the stacked hourglass network model, a single-person upper limb skeleton key point detection model was designed.Deconvolution was employed to replace the up-sampling operation of the hourglass module in the original model, solving the problem of rough feature maps. Integral regression was used to calculate the position coordinates of key points of the skeleton, reducing quantization errors and calculations. Experiments showed that the developed single-person upper limb skeleton key point detection model achieves high accuracy and that the pose estimation method based on the end-to-end approach provides high accuracy and real-time performance.

Peng Gang, Lu Zezao, Chen Shanliang et al.

To address the weak observability of monocular visual-inertial odometers on ground-based mobile robots, this paper proposes a monocular inertial SLAM algorithm combined with wheel speed anomaly detection. The algorithm uses a wheel speed odometer pre-integration method to add the wheel speed measurement to the least-squares problem in a tightly coupled manner. For abnormal motion situations, such as skidding and abduction, this paper adopts the Mecanum mobile chassis control method, based on torque control. This method uses the motion constraint error to estimate the reliability of the wheel speed measurement. At the same time, in order to prevent incorrect chassis speed measurements from negatively influencing robot pose estimation, this paper uses three methods to detect abnormal chassis movement and analyze chassis movement status in real time. When the chassis movement is determined to be abnormal, the wheel odometer pre-integration measurement of the current frame is removed from the state estimation equation, thereby ensuring the accuracy and robustness of the state estimation. Experimental results show that the accuracy and robustness of the method in this paper are better than those of a monocular visual-inertial odometer.

Peng Gang, Lu Zezao, Chen Bocheng et al.

The visual SLAM method is widely used for self-localization and mapping in complex environments. Visual-inertia SLAM, which combines a camera with IMU, can significantly improve the robustness and enable scale weak-visibility, whereas monocular visual SLAM is scale-invisible. For ground mobile robots, the introduction of a wheel speed sensor can solve the scale weak-visible problem and improve the robustness under abnormal conditions. In this thesis, a multi-sensor fusion SLAM algorithm using monocular vision, inertia, and wheel speed measurements is proposed. The sensor measurements are combined in a tightly coupled manner, and a nonlinear optimization method is used to maximize the posterior probability to solve the optimal state estimation. Loop detection and back-end optimization are added to help reduce or even eliminate the cumulative error of the estimated poses, thus ensuring global consistency of the trajectory and map. The wheel odometer pre-integration algorithm, which combines the chassis speed and IMU angular speed, can avoid repeated integration caused by linearization point changes during iterative optimization; state initialization based on the wheel odometer and IMU enables a quick and reliable calculation of the initial state values required by the state estimator in both stationary and moving states. Comparative experiments were carried out in room-scale scenes, building scale scenes, and visual loss scenarios. The results showed that the proposed algorithm has high accuracy, 2.2 m of cumulative error after moving 812 m (0.28%, loopback optimization disabled), strong robustness, and effective localization capability even in the event of sensor loss such as visual loss. The accuracy and robustness of the proposed method are superior to those of monocular visual inertia SLAM and traditional wheel odometers.

Peng Gang, Wei Zeng, Yuri Gordienko et al.

The assessment of energy expenditure in real life is of great importance for monitoring the current physical state of people, especially in work, sport, elderly care, health care, and everyday life even. This work reports about application of some machine learning methods (linear regression, linear discriminant analysis, k-nearest neighbors, decision tree, random forest, Gaussian naive Bayes, support-vector machine) for monitoring energy expenditures in athletes. The classification problem was to predict the known level of the in-exercise loads (in three categories by calories) by the heart rate activity features measured during the short period of time (1 minute only) after training, i.e by features of the post-exercise load. The results obtained shown that the post-exercise heart activity features preserve the information of the in-exercise training loads and allow us to predict their actual in-exercise levels. The best performance can be obtained by the random forest classifier with all 8 heart rate features (micro-averaged area under curve value AUCmicro = 0.87 and macro-averaged one AUCmacro = 0.88) and the k-nearest neighbors classifier with 4 most important heart rate features (AUCmicro = 0.91 and AUCmacro = 0.89). The limitations and perspectives of the ML methods used are outlined, and some practical advices are proposed as to their improvement and implementation for the better prediction of in-exercise energy expenditures.

Nikita Gordienko, Yuriy Kochura, Vlad Taran et al.

Automatic recognition of the historical letters (XI-XVIII centuries) carved on the stoned walls of St.Sophia cathedral in Kyiv (Ukraine) was demonstrated by means of capsule deep learning neural network. It was applied to the image dataset of the carved Glagolitic and Cyrillic letters (CGCL), which was assembled and pre-processed recently for recognition and prediction by machine learning methods (https://www.kaggle.com/yoctoman/graffiti-st-sophia-cathedral-kyiv). CGCL dataset contains >4000 images for glyphs of 34 letters which are hardly recognized by experts even in contrast to notMNIST dataset with the better images of 10 letters taken from different fonts. Despite the much worse quality of CGCL dataset and extremely low number of samples (in comparison to notMNIST dataset) the capsule network model demonstrated much better results than the previously used convolutional neural network (CNN). The validation accuracy (and validation loss) was higher (lower) for capsule network model than for CNN without data augmentation even. The area under curve (AUC) values for receiver operating characteristic (ROC) were also higher for the capsule network model than for CNN model: 0.88-0.93 (capsule network) and 0.50 (CNN) without data augmentation, 0.91-0.95 (capsule network) and 0.51 (CNN) with lossless data augmentation, and similar results of 0.91-0.93 (capsule network) and 0.9 (CNN) in the regime of lossless data augmentation only. The confusion matrixes were much better for capsule network than for CNN model and gave the much lower type I (false positive) and type II (false negative) values in all three regimes of data augmentation. These results supports the previous claims that capsule-like networks allow to reduce error rates not only on MNIST digit dataset, but on the other notMNIST letter dataset and the more complex CGCL handwriting graffiti letter dataset also.

Shiqi Yang, Gang Peng

This paper proposes a novel attention model for semantic segmentation, which aggregates multi-scale and context features to refine prediction. Specifically, the skeleton convolutional neural network framework takes in multiple different scales inputs, by which means the CNN can get representations in different scales. The proposed attention model will handle the features from different scale streams respectively and integrate them. Then location attention branch of the model learns to softly weight the multi-scale features at each pixel location. Moreover, we add an recalibrating branch, parallel to where location attention comes out, to recalibrate the score map per class. We achieve quite competitive results on PASCAL VOC 2012 and ADE20K datasets, which surpass baseline and related works.

Shiqi Yang, Gang Peng

In this paper, we introduce a simple but quite effective recognition framework dubbed D-PCN, aiming at enhancing feature extracting ability of CNN. The framework consists of two parallel CNNs, a discriminator and an extra classifier which takes integrated features from parallel networks and gives final prediction. The discriminator is core which drives parallel networks to focus on different regions and learn complementary representations. The corresponding joint training strategy is introduced which ensures the utilization of discriminator. We validate D-PCN with several CNN models on two benchmark datasets: CIFAR-100 and ImageNet32x32, D-PCN enhances all models. In particular it yields state of the art performance on CIFAR-100 compared with related works. We also conduct visualization experiment on fine-grained Stanford Dogs dataset and verify our motivation. Additionally, we apply D-PCN for segmentation on PASCAL VOC 2012 and also find promotion.